Three-Dimensional Organization of Primary Lens Fiber Cells

Abstract

purpose. To visualize the three-dimensional organization of primary lens fiber cells.

methods. The gene for Green Fluorescent Protein (GFP) was introduced into the lens vesicle using two different vector systems: a replication deficient adenovirus or an expression plasmid. Injected embryos were allowed to develop for several days and then were examined by confocal microscopy.

results. Injection of either vector resulted in GFP expression in primary fiber cells. GFP-expressing cells were heterogeneous in shape and length. Some regions of the fibers were varicose, with diameters >10 μm; regions between the varicosities were often extremely thin, with diameters of <2 μm. No differences in the morphologies of GFP-expressing cells were noted between adenovirus- and plasmid-injected lenses, suggesting that the irregular, undulating, appearance of the primary fibers was not the result of viral infection. Three-dimensional reconstruction of primary fiber cells revealed that, by E6, the posterior tips of the fibers had detached from the lens capsule. The anterior fiber tips remained in contact with the overlying epithelium for 1 to 2 additional days, demonstrating that the formation of the anterior and posterior sutures was asynchronous.

conclusions. The three-dimensional cellular organization of GFP-expressing cells is consistent with previous analyses of fiber cell morphology in the embryonic nucleus of adult human and bovine lenses. The present data confirm that the disorganized appearance of primary fiber cells observed in adult lenses is largely a reflection of developmental processes rather than a consequence of aging.

Primary fiber cells are formed early in embryonic development by elongation of cells at the posterior of the lens vesicle. The elongating primary fiber cells rapidly obliterate the lumen of the lens vesicle. Subsequently, differentiation of cells at the lens equator leads to the formation of secondary fiber cells that come to overlie the primary fibers. Formation of secondary fiber cells occurs continuously throughout life with the original primary fiber cells becoming progressively buried. In the adult lens, the primary fiber cells constitute only a small fraction of the total number of lens cells and are located in the center of the tissue in a region termed the embryonic nucleus. 

Compared to the regular and well-ordered layers of secondary fiber cells, the primary fiber cells are known to be a rather heterogeneous cell population. In human lenses, for example, electron microscopic analysis of lenses revealed that the cross-sectional area of primary fiber cells was 80 ± 68 μm² (mean ± SD) with a range of 7 to 308 μm². ¹ In contrast, the cross-sectional area of secondary fiber cells in the cortex of the lens was found to be much more uniform (24 ± 9μ m²). In another study, light microscopic analysis of bovine lenses also indicated that the cross-sectional profiles of primary fiber cells were very heterogeneous, with an average value of 63 ± 61μ m². ² These data suggest that the cellular morphology and organization of primary fiber cells is very different from that of the secondary fiber cells, which make up the majority of the tissue volume. Although the architecture of the lens nucleus has been studied extensively by scanning electron microscopy, ^{3 4 5} the three-dimensional organization of individual primary fiber cells has not been fully elucidated. Consequently, the cellular features that give rise to the large, observed variations in cross-sectional area have yet to be determined. Here we report a novel approach for visualizing the three-dimensional structural organization of the primary fiber cell population in intact or sliced embryonic lens preparations. The technique involves the use of adenoviral and plasmid expression vectors carrying the gene for Green Fluorescent Protein (GFP), an autofluorescent protein isolated from the jellyfish Aquorea victoria. ⁶  

Materials and Methods

Embryo Injection

Experimental procedures conformed to the ARVO Statement on the Use of Animals in Ophthalmic and Vision Research. Chicken embryos (Truslow Farms, Chestertown, MD) were incubated in a forced draft incubator at 38°C. On day 3.5 of development (E3.5), eggs were removed from the incubator, and a small window was cut in the shell. A single ∼10-nl microinjection was made into the right eye of the embryo using a beveled injection pipette connected to an automatic injector (Drummond Scientific Company, Broomall, PA). The injection was made into the lumen of the lens vesicle as described. ⁷ The injection solution contained either replication deficient adenovirus (2 × 10⁹ plaque-forming units/ml) or plasmid DNA (pDNA; 0.5 mg/ml) and 0.1% Fast Green dye to visualize the injection. After injection, windowed eggs were sealed with tape and returned to the incubator. For comparative purposes, lenses were also obtained from adult White Leghorn chickens (Charles River SPAFAS, Inc., Preston, CT). Adult animals were euthanatized by CO₂ inhalation, and lenses were processed immediately for immunofluorescence (see below). 

Expression Vectors

A replication-deficient adenovirus expression vector (AdCMV5GFP) was obtained from a commercial supplier (Quantum Biotech Inc., Montreal, Canada). AdCMV5GFP contains the GFP gene under the control of an enhanced CMV5 promoter. A plasmid expression vector (pCAGGFP15) containing the potent hybrid CAG promoter ⁸ was constructed in our laboratory. We generated the pCAGGFP15 plasmid by replacing the cytomegalovirus promoter of the pEGFPN1 expression plasmid (Clontech Laboratories Inc., Palo Alto, CA) with the CAG promoter. Preliminary experiments demonstrated that the pCAGGFP15 expression plasmid was more effective at driving GFP expression in primary lens fibers than the pEGFPN1 vector. Before injection, pDNA was purified as previously described. ⁹  

Confocal Imaging

At intervals after injection, embryos were removed from their eggs and examined for GFP expression in the lens. Lenses in which GFP-expressing fiber cells were detected were fixed for 1 hour in 4% paraformaldehyde/phosphate-buffered saline. Fixed lenses were cut into 200-μm midsagittal slices using a Vibratome (model 3000; TPI Inc., St. Louis, MO) as previously described. ¹⁰ Slices were viewed using a confocal microscope (LSM410; Carl Zeiss Inc., Thornwood, NY) equipped with an argon/krypton laser. GFP fluorescence was excited with the 488-nm laser line and detected using a 515 longpass filter. For some samples, a series of optical sections were obtained and reconstructed in three dimensions using software supplied with the instrument. Cross-sectional areas of GFP-expressing fiber cells were calculated from xz projections computed from the three-dimensional reconstructions. As an independent method for visualizing the cross-sectional profiles of primary lens fiber cells, uninjected embryonic lenses were sliced perpendicular to the optical axis and stained with the lipophilic dye, DiOC₆. ¹¹ Adult lenses were fixed, embedded in LR white resin, and sectioned. ² Thin (1 μm) sections were processed for immunofluorescence using an antibody to the major intrinsic protein (MIP) as described. ¹²  

Results

After microinjection of pDNA or adenovirus into the lumen of the lens vesicle, embryos were examined at regular intervals to determine the time course of GFP expression in the lens. GFP fluorescence was first detected on E5, approximately 36 hours after vesicle injection. A typical GFP expression pattern in injected lenses at E5 is shown in Figure 1 . At E5, the lumen of the lens vesicle was filled with primary fiber cells, and the first secondary fiber cells were beginning to differentiate at the lens equator. Injection of pDNA or adenovirus resulted in GFP expression in primary fiber cells located near the optical axis of the lens. The distribution of GFP-expressing fibers differed somewhat between pDNA-injected and adenovirus-injected lenses. In the former, GFP-expressing cells usually formed a tight cluster of cells, whereas in the latter, scattered individual GFP-expressing cells were observed. GFP fluorescence persisted in the central fiber cells at least until E12. Thereafter, there was some decline in intensity, although some lenses exhibited strong fluorescence until E18, the oldest stage examined. 

To better visualize the distribution and morphology of the GFP-expressing fiber cells, midsagittal lens slices were prepared. Viewed in this fashion, it was apparent that GFP expression was almost exclusively restricted to lens fiber cells. Neither pDNA injection nor adenovirus injection resulted in significant GFP expression in the lens epithelium. Occasional GFP-expressing epithelial cells were observed in young (E5–E6) virus-injected lenses (Fig. 2A ), but these were absent at later stages (Fig. 2B) . None of the overlying secondary fibers expressed GFP, suggesting that only those cells initially exposed to pDNA or adenovirus were affected. Although the number and distribution of GFP-expressing fiber cells did not change markedly during development, changes in the morphology of individual cells were noted. The most striking of these was the elaboration of complex membrane processes at later stages. At E5, fiber cells had relatively smooth membrane profiles (inset, Fig. 2A ). However, by E10, the surface of the fibers was marked by the presence of numerous membrane protrusions (inset, Fig. 2B ). These structures presumably correspond to the ball-and-socket interlocking devices observed by SEM. ¹³  

Lens slices, such as those shown in Figure 2 , contain many layers of undisturbed fiber cells. Using the optical sectioning capability of the confocal microscope, we reconstructed the three-dimensional organization of GFP-expressing fiber cells in a virus injected lens at E6 (Fig. 3) . Because adenovirus infected only scattered primary fiber cells (Fig. 1) , these cells were optically isolated in the wild-type background. Viewed in a depth-coded reconstruction, the varicose morphology of the primary fiber cells was apparent (Fig. 3) . Unlike secondary fibers, which are very uniform in shape, GFP-expressing primary fiber cells were an extremely heterogeneous cell population. Each primary fiber consisted of one or more varicosities (>10 μm cross-sectional diameter) joined by thin (< 2 μm cross-sectional diameter) connecting regions. The nuclei were often particularly fluorescent in GFP-expressing fiber cells and were usually located in one of the varicose regions of the cell. 

At E6 (the developmental stage shown in Fig. 3 ) the primary fiber cells retained a connection anteriorly to the lens epithelium. Posteriorly, however, most of the fiber tips were already detached from the lens capsule. The posterior tips of the detached fibers terminated on the lateral membranes of neighboring cells. At these points of contact, the fiber tips were usually swollen into pads (arrows, Fig. 3 ). 

In xz projections, the fiber cells had rounded, cross-sectional profiles. The profiles varied considerably in area depending on whether the section plane passed through a varicose region of the cell (e.g., cell 2 in Fig. 3 ) or a thin connecting region (e.g., cell 1). For the 10 cells shown in the xz projection in Figure 3 , the cross-sectional area was 53.5 ± 58.4μ m² (mean ± SD) with a range of 9μ m² (cell 9) to 199 μm² (cell 2). Cells with similar undulating morphologies also were observed in pDNA injected lenses (data not shown). The three-dimensional reconstructions revealed an irregular primary fiber cell organization. To test whether the normal morphology of these cells had been distorted by the introduction of the exogenous GFP gene, we compared cross-sectional profiles in virus-injected and uninjected lenses. In virus-injected lenses, the cross-sectional profiles were computed from xz projections of the image stack. Uninjected embryonic lenses were sectioned in the equatorial plane and stained with the lipophilic probe DiOC₆ to visualize the plasma membranes. The central region of a lens prepared in this fashion is shown in Figure 4 . A range of rounded, cross-sectional profiles was observed. Examples of large, medium, and small profiles are highlighted in Figure 4 . These profiles were indistinguishable in size and shape from those shown in xz projections of virus infected lenses (Fig. 3) . We also examined cross-sectional profiles of primary fiber cells in the center of adult (1-year-old) lenses using MIP immunofluorescence to visualize the fiber membranes. The size and shape of the cells in the adult lens were indistinguishable from those of the DiOC₆-stained (Fig. 4A) or GFP-expressing (Fig. 3) embryonic primary fibers. 

Discussion

In this study, we described the use of two vectors for introducing GFP into primary fiber cells. In either case, a small volume (∼10 nl) of vector solution was injected into the lumen of the lens vesicle. Injection of pDNA resulted in GFP expression in a cluster of centrally located primary fiber cells. Uptake of externally applied pDNA has been reported in other systems, notably the cornea, where direct application of pDNA to the corneal surface resulted in transfection of corneal epithelial cells. ¹⁴ In the present case, it is possible that a different mechanism may have operated. We have noticed that off-center injection of pDNA into the lens vesicle results in GFP expression in off axis fiber cells (data not shown). This may indicate that only epithelial cells that are physically damaged during the injection procedure take up the pDNA. Primary fiber cells that express GFP subsequently are likely to be derived from these epithelial cells. 

The scattered distribution of GFP-expressing cells in adenovirus-injected lenses was consistent with direct viral infection of cells at the posterior of the lens vesicle. The infection rate was low in the primary fibers and almost undetectable in the anterior epithelium. Interestingly, GFP-expressing epithelial cells were observed more frequently in young lenses (<E6) suggesting that the anterior epithelial cells may be somewhat susceptible to adenoviral infection. Presumably, infected epithelial cells either stop synthesizing GFP at later developmental stages or die, perhaps by virally induced apoptosis. In preliminary studies, we verified the efficacy of the adenovirus vector in HeLa cell cultures. At approximately the same multiplicity of infection as used in lens injections, we observed infection rates of 30% to 40% (data not shown). Thus, the low infection rate observed in the lens fiber cells was not due to lack of virulence in the adenovirus vector. Recently, a cellular adenovirus receptor was cloned from HeLa cells. ¹⁵ The receptor is a 46-kDa cell-surface protein termed CAR (coxsackievirus and adenovirus receptor). The CAR protein mediates high-affinity binding of the adenoviral fiber protein to the cell surface. Cells that lack the CAR receptor, such as ciliated airway epithelia, are resistant to adenovirus infection. ¹⁶ Subsequent to binding to CAR, efficient internalization of the virus also may depend on the presence ofα ₃β₃ orα _vβ₅ integrins. ¹⁷ Thus, the low infection rate of primary fibers cells and the negligible infection rate in epithelia could be due to lack of the CAR receptor and/or the absence of particular integrins. 

An interesting feature of primary fiber cell organization revealed in the present study is the manner in which the fiber tips detach from the capsule. Detachment from the posterior capsule does not appear to be a coordinated event. At E6, some primary fiber cell posterior tips are still attached to the capsule, whereas others (often those of neighboring cells) have detached and terminate on the lateral membranes of adjacent cells. Thus, not all primary fiber cells extend from the anterior to the posterior of the embryonic nucleus. These studies also confirm that the anterior and posterior sutures do not form concomitantly. The formation of the posterior suture precedes that of the anterior suture by 1 to 2 days. 

Injection of virus or pDNA resulted in GFP expression in a limited number of primary fiber cells. This allowed individual GFP-expressing cells to be optically isolated and imaged in a wild-type background. The most striking feature of primary fiber cell morphology revealed by this approach was the heterogeneity of this cell population. In contrast to secondary fiber cells (which have a very uniform morphology), primary fiber cells were shown to be extremely diverse in size and shape. Some regions of the cells were swollen (e.g., the posterior tips of the cells or the portion that accommodated the nucleus); others were extremely thin and tenuous. The morphology of the primary fiber cells was so distinct from that commonly observed in secondary fiber cells that it is appropriate to ask whether the features observed here could be an artifact. Several lines of evidence argue against this. First, the features were observed using both pDNA and viral vectors. Second, cross-sectional profiles computed from xz projections predicted a range of cross-sectional areas that were in excellent agreement with those measured in uninjected embryonic and adult lenses. Finally, the range of cross-sectional areas reported here was similar to that measured in thin sections of primary fiber cells prepared from the lenses of other species. ^{1 2}  

In previous studies of the central region of the adult human lens (the so-called embryonic nucleus), several authors noted the irregular appearance of the primary fiber cells. ^{1 2} In those studies, some fiber cells had very small cross-sectional areas (7μ m²), whereas others were much larger (308μ m²). ¹ Because of the extreme length of the fiber cells, serial reconstruction of mechanically sectioned tissue has not been attempted and consequently, it has not been possible to discriminate between the two models of fiber morphology shown in Figure 4C . The uniform and the varicose models shown in Figure 4C are both consistent with previously published cross-sectional images of primary fiber cells. The data presented here suggest that the varicose model is the more accurate representation. By comparing the morphology of primary fibers in both embryonic and adult lenses, we also have confirmed that an irregular undulating morphology is an inherent feature of the primary fiber cells from the outset. Thus, although the cells of the embryonic nucleus are the oldest fiber cells in the adult lens, their disorganized appearance cannot be attributed to the passage of time. Interestingly, the irregular morphology of the primary fiber cells does not appear to cause an increase in light scattering. Slit lamp examination of the lens nucleus suggests that this is perhaps the most transparent region of the lens. ¹⁶  

We have previously reported the use of cytoplasmically injected expression plasmids for introducing exogenous genes into cortical lens fibers. ⁹ Although the injection technique was effective in secondary fibers, it was not readily applicable to the primary fiber cell population. The use of the vesicle injection technique described in this article complements the earlier study and demonstrates that cloned genes can be targeted to individual cells in geographically distinct regions of the lens. 

 Supported by National Institutes of Health Grants R01-EY09852 (SB), Core Grant EY02687, and an unrestricted grant from Research to Prevent Blindness (RPB) to the Department of Ophthalmology and Visual Sciences. SB is the recipient of an RPB career development award.

 Submitted for publication March 12, 1999; revised September 21, 1999; accepted October 26, 1999.

 Commercial relationships policy: N.

 

Figure 1.

 

Distribution of GFP-expressing cells in E5 lenses after injection of plasmid or adenoviral vectors. Note that, in plasmid injected lenses, GFP-expressing primary fiber cells were tightly clustered near the optical axis (arrow). In virus-injected lenses, scattered primary fiber cells express GFP (arrow). No fluorescence was observed in uninjected control lenses. Scale bar, 250μ m.

Figure 2.

 

GFP expression pattern in adenovirus-injected lenses. (A) At E6, scattered primary fiber cells (arrows) and occasional cells in the anterior epithelium (arrowhead) express GFP. At higher magnification (inset), the GFP-expressing fiber cells are seen to have smooth surfaces. (B) By E10, the GFP-expressing fiber cells (arrow) are buried by layers of newly differentiated secondary fiber cells. At higher magnification, GFP fluorescence is especially strong in elaborate surface structures. Scale bars, 100μ m.

Figure 3.

 

Depth coded three-dimensional reconstruction of a group of GFP-expressing primary fiber cells in an E6 midsaggital lens slice. Optical sections were collected through a thickness of 69 μm. Regions near the top of the image stack (closest to the viewer) are colored red, those near the bottom are colored blue. Ten numbered cells are identified. Note that GFP-expressing fiber cells are extremely diverse in size and shape. The primary fibers are comprised of varicose regions (∗) joined by thin cytoplasmic linkages. The varicose regions often contain the nucleus of the fiber cell (n) although not always (n′). At E6, all fiber cells are in contact with the epithelium anteriorly, but many fibers have detached from the posterior capsule and their posterior tips terminate in thickened pads (arrows) that abut the lateral membranes of adjacent fibers. The variation in primary fiber cell cross-sectional area is shown in the xz projection at the bottom of the figure. The xz projection was computed at the y coordinate indicated by the horizontal yellow line in the xy projection. The numbered profiles correspond to individual GFP-expressing cross-sectioned cells. Scale bar, 50 μm.

Figure 4.

 

Variation in cross-sectional profiles in embryonic and adult primary fiber cells. (A) The central region of an uninjected E17 lens slice cut in the equatorial plane and stained with DiOC₆ to visualize the cell membranes (see text for details). Note the large variation in cross-sectional areas of the fiber cells. Examples of large (blue), medium (green), and small (red) profiles are highlighted. (B) Thin (1 μm) plastic section of primary fibers from the adult lens incubated with anti-MIP to visualize the cell membranes. Note the similarity in size and shape between adult fiber cells and those of the embryo. The membranes of the adult cells appear smoother than in the embryo. However, this is largely explained by the greater z-axis resolution achievable with thin plastic sections. (C) Two possible models of primary fiber cell organization that account for the observed variation in cross-sectional profiles. In the uniform model, the cross-sectional area of a given cell is constant along the length of the cell, but cells of differing diameters are present. In the varicose model, the fiber cells are roughly equivalent in morphology, but the cross-sectional area of a given cell varies along its length. Scale bar, (A, B) 10 μm.

The authors thank Tim Fleming for providing the HeLa cells, Sue Menko for the chicken MIP antibody, and Keith Hruska and Ulises Alvarez for the adult chickens. David Beebe initially suggested the vesicle injection approach, and David Leib drew our attention to the utility of extracellular plasmid injections.