Early lens development in the zebrafish: A three-dimensional time-lapse analysis



In vivo, high-resolution, time-lapse imaging characterized lens development in the zebrafish from 16 to 96 hr postfertilization (hpf). In zebrafish, the lens placode appeared in the head ectoderm, similar to mammals. Delamination of the surface ectoderm resulted in the formation of the lens mass, which progressed to a solid sphere of cells separating from the developing cornea at approximately 24 hpf. A lens vesicle was not observed and apoptosis was not a major factor in separation of the lens from the future cornea. Differentiation of primary fibers began in the lens mass followed by formation of the anterior epithelium after delamination was complete. Secondary fibers differentiated from elongating epithelial cells near the posterior pole. Quantification characterized three stages of lens growth. The study confirmed the advantages of live-cell imaging for three-dimensional quantitative structural characterization of the mechanism(s) responsible for cell differentiation in formation of a transparent, symmetric, and refractile lens. Developmental Dynamics 238:2254–2265, 2009. © 2009 Wiley-Liss, Inc.


The zebrafish provides a rare opportunity to visualize and quantify in real time the microscopic development of the living eye, one of the most exquisite functional systems in biology. In the biological visual system, an effective optical element requires production of a transparent, refractile, collagenous cornea and the differentiation of a symmetric, refractile lens consisting of transparent cells. The synchronized proliferation, migration, and elongation of differentiating lens fiber cells that resist the stress of a changing cellular environment without vasculature, innervation, or lymphatics is regulated by genetic mechanisms that have eluded developmental biologists for decades. While the biological lens is often dismissed as an inanimate optical element similar to an intraocular lens or the synthetic focusing element in a camera, microscope, or telescope, nothing could be more inaccurate. The lens is unique as the only transparent cellular tissue in vertebrates. Microscopic technology for investigating the genetic basis for live-cell differentiation in a vertebrate lens is necessary for characterization of the dynamics of symmetric lens growth and development in vivo. The current study tested the hypothesis that multiphoton imaging using a palmitoylated cyan fluorescent protein transgene as a membrane label in the living zebrafish eye provides the technological advance necessary for in vivo, spatiotemporal analysis of progressive changes in the structural phenotype of differentiating lens fibers from the lens placode to the symmetric, refractile optical lens.

Much work has been performed in vertebrate model organisms to understand human lens development, including in rodent, chick, and Xenopus, which can be difficult for in vivo observation of development. The zebrafish has emerged more recently as an organism with unique advantages over other vertebrate models for studying the visual system (Schmitt and Dowling,1994; Easter and Nicola,1996; Glass and Dahm,2004; Fadool and Dowling,2008). Rapid development outside of a womb, transparency of early embryos, robust reproduction, and relative ease of genetic manipulation allow for experimental studies in vivo that would be impractical in mammalian species. One of the greatest advantages of the zebrafish is the ability to perform real-time cell imaging of living, developing embryos, which avoids the problems of distortion and artifacts caused by ex vivo or in vitro studies. These benefits, combined with the evolutionary development of an advanced visual system with many similarities to humans, make the zebrafish an ideal vertebrate model organism for live-cell research on the eye.

The vertebrate lens begins as a thickening of the surface ectoderm to form a lens placode (Lang,2004; Lovicu and Robinson,2004; Vihtelic,2008; Schoenwolf and Larsen,2009). The mammalian and avian lens placode undergoes invagination and pinches off from the surface ectoderm as a hollow lens vesicle. Cells in the posterior half of the vesicle elongate to fill the vesicle cavity and differentiate into primary fiber cells. Cells in the anterior half of the vesicle become the anterior epithelium of the lens. Epithelial cells proliferate and migrate peripherally to the equator, where coordinated elongation and differentiation of transparent secondary lens fibers begins. The elongating fibers detach from the lens capsule and form layers of secondary fibers surrounding the embryonic primary fibers in the lens nucleus. Synchronized development of the three lens cell types, i.e., anterior epithelium, primary fibers, and secondary fibers, is imperative for the formation of a transparent, symmetric, refractile lens that has similar properties in adult zebrafish and adult mammals (Dahm et al.,2007; Greiling and Clark,2008).

While a few impressive histological studies have illustrated the early development of the zebrafish lens (Glass and Dahm,2004; Soules and Link,2005; Dahm et al.,2007), many details need elucidation, and the zebrafish lens has some important differences from the human lens during early development. Instead of invagination and vesicle formation, the zebrafish lens develops as a solid mass of cells that delaminates from the surface ectoderm (Schmitt and Dowling,1994; Easter and Nicola,1996). The process of development and differentiation of lens epithelium, primary fibers, and secondary fibers without a lens vesicle stage remains poorly understood. An in-depth understanding of the normal structural phenotype during development will be valuable for characterization of the multitude of lens mutants created by forward genetic screens and morpholino injection, and for advancing our knowledge of the fundamental processes required for human lens development. Live developmental imaging will permit the analysis of subtle phenotypic changes created by gene mutants and morphants that more closely mimic human disease and aging and could be missed easily by distortions introduced using conventional histological sectioning.

This report is the first systematic, high-resolution, quantitative, real-time, three-dimensional (3D) description of lens development in a living vertebrate from the placode stage at 16 hours postfertilization (hpf) through 4 days postfertilization (dpf). Live-cell imaging provided a direct and progressive view of the establishment of primary fibers, an anterior epithelium, and secondary fibers during development of the symmetric, transparent lens. Ectodermal cells in the cephalic placode thickened, proliferated, and reorganized into the primary fibers of the developing lens core without formation of a vesicle stage. No evidence for apoptosis as a mechanism for delamination and separation of the zebrafish lens from the surface ectoderm was observed. Production of new, secondary fibers expanded the lens diameter with growth of the zebrafish eye. The observation of the development of primary lens fibers and the proliferation, migration, and elongation of secondary fibers offered a direct view in a living vertebrate animal of the process of differentiation, which included the fundamental stages of an epithelial–mesenchymal transition and had many similarities with human lens development. The study confirmed the advantages of live-cell imaging and will pave the way for future analysis of structural studies of the mechanisms responsible for cellular differentiation in formation of a transparent, symmetric, and refractile lens.


Time-lapse, in vivo, high-resolution imaging characterized the cellular composition and behavior of the lens at each major stage of development.

Lens Placode

At 16 hpf, the basal-to-apical height of the cells of the surface ectoderm approximately doubled as the simple cuboidal epithelium of the surface ectoderm began to form a columnar epithelium in the lens placode, the region overlying the center of the retinal anlage (Fig. 1; Supp. Movie. S1, which is available online). When viewed from the surface, the lens placode was circular, approximately 8 cells in diameter, and was composed of columnar cells that were relatively uniform in size and shape. Not all cells in contact with the developing retina became columnar: 2–3 cells at the border of the placode remained cuboidal in shape, although these cells were also closely apposed to the developing retina. The retina was visible as a solid optic primordium in which the future neural retina and future pigment epithelium each formed single layers of pseudostratified columnar epithelium that were tightly apposed with no lumen visible. In the zebrafish, the lens placode at 16 hpf resembled the mammalian or avian placode.

Figure 1.

At 16 hours postfertilization (hpf), lens placode. A: Axial section through the center of the lens and retina. A single layer of surface ectoderm formed the lens placode overlying and closely apposed to the outer layer of the retinal anlage. B: Cells of the lens placode were colored digitally and oriented in an axial section corresponding to A. In the image, the basal surfaces of the cells are down and the apical surfaces are up. The cells are consistent in height and vertically or obliquely oriented. C: In a three-dimensional surface view of the head ectoderm, the circular arrangement of the colored cells is observed. See Supp. Movie S1 for 3D view.

By 18 hpf, many of the cells in the lens placode more than doubled in height from 16 hpf and appeared as a mass of cells organized as a flattened spheroid (Fig. 2; Supp. Movie S2). The curvature at the posterior surface of the lens mass was greater than at the anterior surface. The lens mass appeared to be radially symmetric when viewed from the surface, although it lacked the spherical symmetry found in a functional adult lens. The lens mass was two or three cell-layers thick at the center with a single layer remaining laterally. It was unclear whether the additional cells comprising the placode were formed through proliferation between 16 and 18 hpf or whether cells had been recruited from the surrounding surface ectoderm or both. At the surface, an abrupt change in morphology made elongated cells of the lens placode clearly distinguishable from cuboidal cells of the surface ectoderm. The lower surface of the lens mass was approximately 14 cells in diameter. These cells were more than twice as tall as the cells at 16 hpf and were oriented at an oblique angle toward the center of the lens mass. A second, morphologically distinct group of cells in the anteromedial region were shorter and more rounded than the elongated cells at the posterior–lateral lens border. A cluster of small, round cells were located just anterior to the surface of the center of the lens mass and appeared to be apoptotic bodies. These cells were probably held in place by the overlying periderm (which did not label with membrane-tagged cyan fluorescent protein [mCFP]). The retina remained in close apposition to the posterior–middle of the lens mass. A small vitreal space (<5 μm) was present posterior to the lateral sides of the lens.

Figure 2.

At 18 hours postfertilization (hpf), lens placode. A: Axial section through the center of the lens mass with the retina digitally removed. The width and thickness of the placode increased as the cells elongated and formed a posterior layer of differentiating cells deep to undifferentiated round cells filling the top/center of the lens mass. Inset: orientation of axial sections taken through lens mass in A–E. BE: Images oriented as in A at selected sections through the lens mass and retina. F: Axial view of a three-dimensional (3D) projection of the lens mass with central cells colored digitally. Small, round apoptotic bodies (blue) were present at on the top surface of the lens mass. G: Oblique view of a 3D projection of the lens mass. H: Surface view of 3D projection of lens mass showing circular shape of the lens mass. See Supp. Movie S2 for 3D view.

Delamination and Development of the Lens Mass

By 20 hpf, the lens mass had elongated perpendicular to the surface ectoderm and narrowed in the equatorial dimension (Fig. 3, Supp. Movie S3). The dramatic change in shape indicated a significant reorganization of the cells comprising the lens mass. At this stage, three distinct cell morphologies were observed: (1) Cells at the posterior and lateral surfaces were tall and similar in height to the elongated cells at 18 hpf. These cells formed an ordered single layer that established a smooth, regular posterior and lateral border of the developing lens. These cells appeared to have well-defined basal and apical surfaces with the basal cell surface 2–3 times wider than the apical surface. (2) Cells in the center of the elongated lens mass were round or ovoid shaped (like cells at the anterior border in the 18 hpf lens) and irregularly clustered in the center of the lens stalk. (3) Cells at the anterior lens border in contact with the surface epithelium were elongated although the orientation was more parallel to the surface ectoderm and irregularly arranged. The lateral transition between the anterior cells of the lens mass and flat cells of the surface ectoderm remained distinct. Small, round bodies that resembled apoptotic cells were observed at the anterior surface of the lens. The posterior border of the lens mass remained in contact with the deepening optic cup. Additional round, apoptotic-like cells were visible in the vitreal space separating the lateral lens mass from the anterior margins of the developing retina, and it was unknown whether these cells originated from the lens, retina, or a different tissue. No lens vesicle was present.

Figure 3.

At 20 hours postfertilization (hpf), lens mass formation and beginning of delamination. A: Axial section through the center of the lens mass and the stalk connecting to the surface ectoderm; retina digitally removed. The thickness of the lens mass increased, bulging toward the retina, and elongated, fiber-like cells appeared along the deep and lateral boundaries of the lens mass surrounding a central core (stalk) of round, undifferentiated cells. Inset: orientation of axial sections taken through lens mass and stalk in A–F. BE: Images of the same selected sections through the lens mass, stalk, and retina. F: Axial view of a three-dimensional (3D) projection similar to Figure A with the central round and oval cells colored digitally. Round apoptotic bodies (blue) were present at the top surface of the lens mass and between the lens and retina. G: Axial view of 3D projection of lens mass and stalk showing colored central cells and apoptotic bodies. H: Surface view of G in which the colored cells of the central stalk were surrounded by elongating fiber-like cells. See Supp. Movie S3 for 3D view.

Delamination and Detachment From the Surface Ectoderm

At 22 hpf, the maximum width of the lens mass at the equator was approximately the same as at 20 hpf, while the cells attached to the surface ectoderm narrowed to form a stalk two to three cells wide connecting the developing cornea with the developing lens (Fig. 4; Supp. Movie S4). The shape of the lens mass became round with very smooth, regular posterior and lateral borders all the way to the cellular connection with the surface ectoderm. Three morphologically distinct cell types were observed in the lens: (1) The posterior and lateral borders of the lens were formed by a single-layer of tall columnar cells with wider basal than apical cell surfaces. These cells radiated out from the central core. (2) A cluster of cells in the middle of the lens appeared tear-drop shaped with their narrow edges facing toward the center in a more organized arrangement than at 20 hpf. (3) Cells at the anterior–middle of the lens and surface ectoderm were rounded or cuboidal in shape with an irregular arrangement. These three cell types appeared to correspond to the primary fiber cells, the embryonic nucleus, and undifferentiated cells of the original lens mass. A few cells were transitional between the anterior lens and the flat cells of the surface ectoderm, and it was unclear whether these cells would ultimately become part of the lens during delamination or remain part of the surface ectoderm and contribute to the developing corneal epithelium. Only a small number of apoptotic-appearing cells remained and were visible anterior to the lens, while no apoptotic cells were visible between the lens and retina.

Figure 4.

At 22 hours postfertilization (hpf), delaminating lens mass. A: Axial section through the center of the delaminating lens; retina digitally removed. Differentiating fiber cells in the lens mass surrounded undifferentiated central cells and appeared to radiate out from central cells as the lens mass detached from the surface ectoderm. A few cells in the remaining central stalk attached the developing lens to the future cornea. Inset: orientation of axial sections taken through lens mass and stalk in A–C,E,F. B,C,E,F: Axial sections through the lens and retina. D: In the axial view, central round or ovoid cells are colored digitally. G: Elongated primary fiber-like cells (light green) radiated out from the central core of cells in the lens mass. See Supp. Movie S4 for 3D view.

By 23 hpf, the lens mass remained attached to the surface ectoderm by a stalk that was only one cell wide (Fig. 5, Supp. Movie S5). The surface ectoderm was a regular, single-layer of flat, cuboidal epithelium above the lens except where it was still attached to the lens in the center. The lens mass began to take on a spherical shape, although the posterior border appeared flat when compared with 22 hpf. Cells at the lateral boundary of the lens remained columnar and fiber-cell-like with widened basal surfaces as if migrating along a developing capsule. Rather than radiating out from the central core as at 22 hpf, the fiber-like cells began to curve and appeared to wrap around the central core. Cells in the central core of the lens mass remained large and tear-drop shaped with their narrow edges facing the center of the lens, and cells at the anterior–middle of the lens remained rounded or cuboidal in shape in an irregular arrangement. No lens vesicle was observed and no apoptosis was visible in the cells at the periphery of the lens mass.

Figure 5.

At 23 hours postfertilization (hpf), final detachment. A: Axial section through the lens center; retina digitally removed. Elongating primary fiber cells began to radiate out and surround the core of round cells. The lens remained connected to the developing cornea by a narrow stalk, only one cell thick. Inset: orientation of axial sections taken through the lens in A–C,E,F. B,C,E,F: Axial sections through the lens and retina. D: Axial three-dimensional (3D) projection view of the lens and cornea with central cells in the lens mass colored digitally. G: Magnification of the connecting stalk of colored cells in the lens. The stalk consisted of one cell in the developing cornea (purple) connected to one cell in the developing lens (red). See Supp. Movie S5 for 3D view.

At approximately 24 hpf, the developing lens mass separated completely from the surface ectoderm, which remained a continuous single layer of epithelial cells at the surface of the head of the fish (Fig. 6, 7, Supp. Movies S6–7). Cells in the posterior–middle of the developing lens continued to enlarge and take on a rounded shape, forming a nuclear “organizing center” around which primary fiber cells elongated and migrated in an ordered but asymmetric manner. Lateral columnar cells observed at 23 hpf appeared to elongate further by 24 hpf to form arcs or layers of primary lens fibers surrounding the developing nucleus. Cells at the anterior border of the delaminated lens began to organize into a single-layer of epithelium. Cells deep to the developing anterior epithelium still appeared disorganized and undifferentiated. The lens was separated from the apical surface of the optic cup by only a few micrometers of vitreal space. No lens vesicle was present.

Figure 6.

At 24 hours postfertilization (hpf), after detachment. A: Axial section through the lens center and cornea; retina digitally removed. Primary fibers continued to elongate and surround the remaining oval cells in the center of the spherical lens that has separated from the cornea. Inset: Orientation of axial sections taken through the lens in A–C,E,F. B,C,E,F: Axial sections through the lens, retina, and cornea. See Supp. Movie S6 for 3D view. D: Axial projection view of lens and cornea in which the central cells are colored digitally. G: Oblique view of D in which the central cells are surrounded by primary fibers.

Figure 7.

At 24 to 37 hours postfertilization (hpf), lens development; retina digitally removed from images. A–C: Axial section through the lens center. A′–C′: Equatorial sections through the lens center. A: At 24 hpf, axial section. Primary fiber cells surround the organizing center. An anterior epithelium was starting to form. A′: At 26 hpf, lens equatorial section of the central core surrounded by primary fiber cells. The anterior epithelium surrounded the primary fiber cells. B: Axial section; at 28 hpf, the lens anterior epithelium was complete. A few round cells remained in the lens center surrounded by differentiating fibers deep to the epithelium. B′: At 29 hpf, equatorial section at a similar stage to B. C: The 36 hpf axially oriented lens contained central cells surrounded by layers of secondary fibers that appeared to be formed from the epithelium. C′: At 37 hpf, equatorial section at a similar stage to that in C in which there is the central core of cells surrounded by primary fibers and an external layer of epithelial cells. The white bar represents 20 μm; all images are to the same scale. See Supp. Movies S7A S7A′, S7B, S7B′, S7C, S7C′ for 3D views.

Relative changes in cell shape during delamination suggested that proliferation, migration, and adhesion were responsible for a gradual remodeling of the lens mass that resulted in a separate lens and cornea at approximately 24 hpf. The morphology of the cells that connected the lens to the surface ectoderm during delamination suggested that apoptosis was not the main factor in separation of the lens and cornea. A few cells that appeared apoptotic were visible at the periphery of the lens from 18 to 22 hpf and the origin of these cells was unclear. There may have been apoptosis occurring in the undifferentiated cells deep to the developing anterior epithelium at 24 hpf.


By 28 hpf, the morphology of the differentiated lens cells appeared to represent the cell types expected in the adult lens: (1) an epithelium was present that was not limited to surrounding the anterior half of the lens; (2) primary fiber cells wrapped around the large round cells in the core of the lens nucleus; and (3) secondary fiber cells appeared to be elongating and migrating from a developing bow region posterior to the lens equator (Fig. 7; Supp. Movies S7B, S7B′). Cells at the anterior surface formed a single-layer of tall, cuboidal epithelium, which covered the entire anterior hemisphere of the lens and extended posteriorly beyond the lens equator as evidenced by its organized appearance in the equatorial section in Figure 7B′. This indicated that the developing bow region in the zebrafish was located more posterior than the bow region in the mammalian lens. The core of the lens nucleus was composed of a few large, round cells. The number of layers of fiber cells around the core of the nucleus had increased from 24 hpf, although these elongating cells could have originated from a developing bow region and/or from proliferation of primary fibers in the lens mass. The apical surface of the anterior epithelium was in direct contact with underlying fibers except where a few undifferentiated cells remained at the anterior pole just deep to the epithelial layer. It was presumed that most of the undifferentiated cells that were present at 24 hpf between the anterior epithelium and the primary fibers had either differentiated into epithelium or undergone apoptosis by 28 hpf. The lens appeared spherical, and the overall diameter had increased a small amount.


The 36 hpf lens was similar in overall structure to the 28 hpf lens, although newly added fiber cells were smaller and more compact than at 28 hpf (Fig. 7; Supp. Movies S7C, 7C′). The height of the cuboidal cells comprising the anterior epithelium decreased by approximately half, and the epithelium appeared to extend posteriorly past the lens equator. Layers of lens fibers were added symmetrically around the lens, although an umbilical suture was not yet obvious. The lens looked spherical in the equatorial dimension and lentoid in the anterior–posterior dimension. Cell membranes lost their smooth appearance and looked jagged between the central cells as if establishing shallow interdigitations between cells. Cells of the developing corneal endothelium were visible migrating toward the center of the pupil but had not yet met in the middle. Only a single layer of epithelial cells was observed in the center of the developing cornea.

At 48 hpf, the lens had increased in size and appeared spherical at the equator and lentoid in the anterior–posterior direction (Fig. 8, Supp. Movies S8A, S8A′). Cells of the anterior epithelium continued to decrease in relative height and formed a flat cuboidal epithelium with less mCFP expression. Fiber cells in the cortex were narrow and elongated, making it difficult to see cytoplasm between the cell membranes. The posterior tips of newly added elongating fiber cells appeared to meet at the midline establishing a posterior suture. Borders between cells looked increasingly jagged and there was now a “blurry” shell of cells in the outer cortex with membranes that were not defined by a mCFP outline, although the central cells were still clearly defined. Two complete cell-layers were visible at the cornea, which would be expected as a layer of epithelium and a layer of endothelium were being established. The distance between the posterior lens and the retina (vitreous space) increased to approximately 10 micrometers except at the anterior margins of the retina, which remained in contact with the lens.

Figure 8.

At 48 to 97 hours postfertilization (hpf), lens development; retina digitally removed from images. A: At 48 hpf, axial lens. A posterior umbilical suture first appeared. A′: Equatorial section. At 49 hpf, flat epithelial cells had decreased membrane-tagged cyan fluorescent protein (mCFP) expression. B: Axial section. At 72 hpf, individual cells were difficult to distinguish in the lens nucleus. B′: Equatorial section. At 73 hpf, the fluorescence in the nucleus was homogeneous and intense, surrounded by layers of primary and secondary fibers. C: Axial section. At 96 hpf, a shadow from the retina obscured part of the lens. C′: Equatorial section. At 97 hpf, elongated secondary fibers were organized into symmetric radial layers similar to adult lenses in the mammal or avian. The white bar represents 20 μm; all images are to the same scale. See Supp. Movies S8A, S8A′, S8B, S8B′, S8C, S8C′ for 3D views.

At 3dpf, when the zebrafish embryos would have hatched into larvae, the width of the lens increased at the equator and the shape was spherical in all dimensions, suggesting a functional optical element in the visual pathway (Fig. 8, Supp. Movies S8B, S8B′). An umbilical suture was visible at both the anterior and posterior poles. Individual cells could not be distinguished in the entire lens nucleus, which looked like a single bright sphere. This was not due to overexposure of the fluorescence signal, which appeared brighter in the nucleus than the well-defined cortical cells, because it was not altered by increasing or decreasing the contrast of the image. The loss of mCFP definition may correlate with primary fiber cell maturation and organelle break-down in the lens nucleus. The anterior epithelium was only clearly visible when the image contrast was increased threefold. The epithelial cells formed a flat layer that extended posteriorly to the developing bow/transitional zone. Newly added fiber cells formed radial columns, which were visible in the equatorial view, similar to the chick lens which also forms an umbilical suture (Beebe et al.,2001). Because the size of the retina had increased and blocked the penetration of the 2-photon laser to the lens, a shadow from the retina made the posterior half of the lens dimmer than the anterior. Two distinct cells layers were visible at the cornea, although mCFP expression was dimmer than in the lens or retina.

By 4 dpf, the zebrafish lens was a large spherical version of the 3 dpf lens, which increased in size by additional layers of secondary fiber cells (Fig. 8, Supp. Movies S8C, S8C′). The bright central core of cells that lacked mCFP definition also increased in size. The retinal shadow made it very difficult to clearly image the posterior lens.

Lens Dimensions

Quantification of the growth of the lens confirmed the descriptive results observed using 3D multiphoton microscopy of the living lens. The maximum equatorial diameter of the lens was measured at selected time points from 16 hpf to 40 hpf in 14 different zebrafish (Fig. 9, solid lines). Three stages of growth were observed. In stage one (16 hpf to 18 hpf) the width of the lens placode/lens mass increased rapidly, while the overall shape was a bulging disc. In stage two (18 hpf to 21 hpf) the lens mass increased in the anterior–posterior dimension, and the equatorial width of the lens decreased by approximately 30%, suggesting a dramatic cellular rearrangement in the lens mass due to cell migration and changes in cell shape required for establishment of a spherical lens. This tissue rearrangement is analogous to invagination of the lens cup in mammals to establish the spherical shape of the hollow lens vesicle. In zebrafish, no vesicle was observed and the developing lens remained a solid mass of cells. In stage three, the equatorial diameter of the lens mass grew linearly at roughly 1μm/hr during the 21 hpf to 40 hpf time period.

Figure 9.

Lens dimensions with age. The maximum diameter of the lens equator in micrometers (solid lines) and width of the lens stalk in micrometers (dashed lines) were plotted as a function of age (hpf). Three stages of development were observed on the basis of quantification of growth with age. From 16 to 18 hpf, the diameter of the lens placode increased 15–20%. From 18 to 22 hpf, the equatorial diameter decreased approximately 30% during formation of the lens mass. The decrease corresponded with cell migration and increased elongation in the anterior–posterior dimensions. From approximately 19 to 25 hpf, the width of the lens stalk connecting the lens to the cornea (dashed lines) gradually decreased until separation of the lens the developing cornea was completed. At this stage, a slow progressive increase in diameter of the spherical lens was observed, which correlated with the addition of lens fibers. The quantitative measurements of lens dimensions were consistent with the expansion of the placode followed by delamination of the lens mass, which involved cell proliferation, migration, and elongation, the same processes necessary for lens development in mammals. Each line represented measurements from one fish of the lens placode, lens mass, and spherical lens as the diameter changed with time of development. Slight variations between individual lenses were expected and were observed.

At approximately 19–20 hpf, just before the equatorial diameter of the spherical lens began to increase progressively, the connection to the surface ectoderm decreased by approximately 20μm/hr or 1-cell per hour until the progressive delamination of the developing lens from the developing cornea was complete at 24 to 25 hpf (Fig. 9, dashed lines). Some variability was observed in the time of separation/delamination, which may have been related to the extended time that the embryos were mounted in agarose. In summary, the developing lens was transformed from a 2D placode in the surface ectoderm of the zebrafish head to a 3D, spherical optical element detached from the future cornea in a process that involved coordinated changes in lens shape and size. While the mechanism for establishment of the three major cell types is different in the zebrafish and the mammal, the resulting optical element is symmetric, transparent, refractile, and resembles organization in the mammal.


Although the structure and function of the adult zebrafish lens resembles the mature human lens with respect to symmetry, refraction, transparency, and optical function, a different mechanism of early structural development was observed. The zebrafish lens began as a lens placode, similar to mammals, and through progressive delamination rounded up to form a spherical solid mass of cells which separated from the developing cornea to form a distinct lens by 24 hr postfertilization. A lens vesicle which characterizes the mammalian lens was not observed.

A summary diagram of the differentiation of live cells in the living vertebrates lens characterized three cell types on the basis of 3D live-cell imaging (Fig. 10). The lens placode formed at 16 hpf as a layer of columnar cells (purple membranes), which elongated vertically into developing primary fibers (blue membranes) by 18 hpf. The disc-shaped lens mass began to round up by 20 hpf with a change in the morphology of the differentiating primary fiber cells radiating posteriorly and laterally from a central cluster of rounded cells. The elongating primary lens fibers wrapped around the cluster of round cells in the core of the embryonic lens nucleus (red membranes). Detachment of the lens mass from the cornea (orange membranes) occurred over a period of approximately 19–24 hpf. During this time period, cells at the anterior surface of the lens mass (yellow membranes) reorganized and appeared to lose adhesion contacts with the surface ectoderm. Apoptosis was not obvious in these cells. After detachment, cells at the anterior surface of the lens formed a single-layer of cuboidal epithelium (green membranes). A group of morphologically undifferentiated cells deep to the anterior epithelium (yellow membranes) at 24 hpf presumably underwent apoptosis or differentiated so that by 36 hpf, the anterior epithelium covered approximately 7/8 of the lens surface and was in direct contact with the underlying fiber cells. Once the anterior epithelium was established, secondary fiber cells elongated and migrated from a transitional region located posterior to the equator. The posterior suture was visible at 48 hpf and formed before the anterior suture, which was similar to chick embryonic lens development in which the posterior suture forms at embryonic day 6, 1–2 days before the anterior suture (Shestopalov and Bassnett,2000). The complete umbilical suture was obvious by 3dpf, when the lens became a functional optical element in the visual pathway (Easter and Nicola,1996).

Figure 10.

Summary diagram representing the progressive stages of lens development in the zebrafish from the placode to spherical lens and the presumptive cellular differentiation at each stage. The corresponding three-dimensional (3D) movie of each 2D image can be referenced for more detail. Cell membranes have been pseudocolored. A: At approximately16 hours postfertilization (hpf), the cells in the surface ectoderm are orange, and the cells in the lens placode are purple. See Supp. Movie S1. B: At 18 hpf, elongating fiber-like cells are blue. See Supp. Movie S2. C: At 20 hpf, the cells in the organizing center (red) of the delaminating and elongated lens mass are surrounded by columnar primary fiber cells (blue). See Supp. Movie S3. DF: At 22 hpf, 23 hpf, and 24 hpf, respectively, delamination completes the separation of the spherical lens from the developing cornea. Morphologically undifferentiated cells at the anterior surface of the lens mass (yellow) appeared to reorganize and lose adhesion contacts with the surface ectoderm. See Supp. Movies S4–S6. G: After delamination, at approximately 24 hpf, the anterior epithelium (green) began to form. See Supp. Movie S7A. H: At 28 hpf, a transitional zone corresponding to the bow region in the mammal was located on the posterior half of the lens. See Supp. Movie S7B. I: By approximately 36 hpf, a complete anterior epithelium (green) was in direct contact with the underlying fiber cells (blue). An umbilical suture appeared at the posterior pole and the transitional region was observed laterally where the epithelium was found. See Supp. Movie S7C. The white bar in A represents 20 μm; all images to same scale.

The formation of primary fiber cells was one of the first differences between zebrafish and mammalian lens development. Mammalian primary fibers are posterior epithelial cells that elongate to obliterate the lens vesicle. Once elongation begins, primary fiber cells exit the cell cycle (McAvoy,1978; Lovicu and McAvoy,1999). In the absence of a lens vesicle, primary fiber formation needs to be redefined. The results of live-cell imaging determined that primary fibers in the zebrafish lens originated in the lens placode and reorganized to form layers surrounding the cells in the embryonic lens core or “organizing center.” It is unknown when these cells exit the cell cycle. Live-cell imaging determined that secondary fibers began to form from the transition zone at approximately 28–36 hpf, so all fibers formed before that stage were primary fibers (which did not differentiate from an anterior epithelium). The organization of the primary fibers in the zebrafish lens nucleus was quite different from that of the mammalian embryonic lens nucleus.

In contrast, secondary fiber cell formation and morphology was similar in the zebrafish and the mammal. In zebrafish, secondary lens fibers formed from the proliferative zone of the anterior epithelium and migrated toward the posterior pole of the lens where they elongated to form semicircular layers surrounding the primary fibers. Secondary fibers elongated and migrated from a transitional zone in an organized, symmetric manner to form an umbilical suture. Like the adult zebrafish lens (Dahm et al.,2007), the position of the transition/bow region in the embryonic zebrafish lens was located posterior to the equator rather than at the equator in the mammal. The result of a posterior transition zone was that the anterior elongation of secondary fibers was greater than the posterior elongation, which could account for the appearance of the posterior suture before the anterior suture. It is unknown whether or not the proliferative zone of the anterior epithelium is also located more posterior than in the mammal.

The most obvious difference between the mammalian and zebrafish embryonic lens was the lack of a vesicle stage. It is unclear whether or not this is unique to the zebrafish. It has been proposed that other fish and amphibian species also develop solid lenses rather than lens vesicles (Dahm et al.,2007). Within zebrafish organogenesis, the lack of a vesicle stage is not unique. The neural “tube,” optic “vesicle,” and otic “vesicle” form as solid masses of cells and later undergo cavitation to develop lumina (Schmitt and Dowling,1994; Haddon and Lewis,1996; Lowery and Sive,2004). Even earlier in embryogenesis, the zebrafish blastocyst lacks a blastocoele and the gastrula lacks an archenteron (Kimmel et al.,1995). A similar genetic mechanism may be responsible for the lack of cavity formation which is a common feature of organogenesis during zebrafish development.

The process of detachment of the lens mass from the surface ectoderm is another difference between zebrafish and mammalian lens development. Previous reports indicated that the separation of the lens mass involved apoptosis of cells between 24–26 hpf (Soules and Link,2005; Dahm et al.,2007). The current in vivo study determined that separation of the lens and cornea was an incremental process that occurred between 19 hpf and 24 hpf as the lens rounded up and the stalk connecting the lens to the cornea gradually decreased in width, presumably through changes in adhesion contacts until division of the two tissues was complete. Apoptotic cells were not visible in the cell layers/stalk directly connecting the lens and cornea. This process seemed analogous to invagination and pinching off of the lens vesicle in mammals, even though the zebrafish lens mass is a solid ball of cells that round up and pinch off. Additional studies are needed to understand the genetic mechanisms that control cell adhesion, shape, and migration during separation of the lens and cornea.

Although apoptotic cells were not seen between the lens and cornea, multiple small apoptotic bodies were observed anterior to the lens surface from 18–22 hpf, presumably held in place by the periderm, which covers the embryo from 3.5 hpf until at least 48 hpf and is probably shed around the time the embryo hatches (Kimmel et al.,1990). The periderm is a specialized superficial epithelium derived from the enveloping layer that does not contribute to any adult tissue in the mammal, and does not express mCFP (which is driven by a Pax6 enhancer) like the head surface ectoderm and retina. The close relationship between the apoptotic bodies and the surface of the developing lens could suggest that apoptosis is occurring in the lens from 18–22 hpf. Apoptosis appeared to occur in cells deep to the anterior epithelium from 23–28 hpf. Further experiments are needed to confirm this hypothesis and elucidate the role of apoptosis in zebrafish lens development.

Quantitative results of the dynamic growth of the lens or other tissues can only be achieved reliably using live-cell imaging, which is free from the well-known problems of shrinkage and tissue distortion that occur with traditional fixation and embedding procedures. The advantage of the zebrafish for live-cell imaging is the rapid time course from lens placode to a symmetric, transparent, refractile lens. Only live-cell imaging permits quantification of the dynamics of differentiation, development, and growth in real time. In the zebrafish lens, three components were observed in the growth curve. One component correlated with the expansion of the placode during the first few hours of development. With continued growth, a second component correlated with delamination from the surface ectoderm to form the lens mass. The third component of the growth curve correlated with slow progressive growth of the functional lens. Quantitative growth curves will be useful for the evaluation of specific molecular and genetic mechanisms, which may have subtle effects on normal differentiation and growth of the developing transparent lens in the zebrafish, an important advance over descriptive methods in the analysis of genotype–phenotype relationships.

A major challenge in biology is quantifying relationships between the structural variations and cellular adaptations in tissue growth and development regulated by gene expression. To date, genetic and molecular approaches have been at a loss to define the mechanisms regulating the dynamic variations in the differentiation that account for the structural specializations and symmetry of the transparent and refractile lens cell structure. Currently, the refracton hypothesis is the singular theory addressing the molecular genetics for transparency and refractive power of biological optics (Piatigorsky,2001,2008). The cornerstone of the refracton hypothesis is that the lens and cornea accumulated diverse, taxon-specific multifunctional proteins by a gene sharing strategy, which contributed to the optical properties of both tissues. Characterization of their common genetic mechanisms, such as gene sharing, can be expected to expand the understanding of fundamental genetic processes contributing to transparency in biological cells and tissues. The zebrafish has unique experimental advantages that make it an excellent model for study and quantification of the development of the transparent lens. The capability to conduct quantitative evaluations of development in a living system is a dramatic advance that can benefit the study of genotype-phenotype relationships where genetic modification can be subtle and difficult to characterize. Studies of cell migration, adhesion, polarity, and coordinated cellular differentiation that would be impossible or extremely difficult in a living mammal can be done in the zebrafish. The preliminary quantitative structural studies presented in this report begin the systematic correlation of phenotypic reorganization with development to help elucidate the fundamental mechanisms responsible for the establishment and maintenance of a symmetric, refractile, transparent optical element in the visual system of a living vertebrate animal.



All imaging was done with the Q01 transgenic fish line, donated from the laboratory of Dr. R.O.L. Wong. Q01 fish ubiquitously express cyan fluorescent protein fused to a membrane-targeting sequence of zebrafish Gap43 (mCFP), driven by an EF1α promoter and a hexamer of the DF4 pax6 enhancer element (Godinho et al.,2005). Q01 fish were generated in a wild-type background and develop and reproduce normally. Adult fish were housed at 28.5°C in accordance with the University of Washington Institutional Animal Care and Use Committee.

Images and videos in Figures 1–8 and Figure 10 were taken from three different representative zebrafish embryos. Any observations made on a single animal have some likelihood of being anomalous or unrepresentative of the development of a structure within a population. At least 20 and as many as 40 embryos at each time point were observed under identical conditions, and the images and videos presented were representative of the normal 3D morphology at each stage of development. Fish “A” was followed from 16 hpf to 24 hpf and images were taken through the optical axis; fish “B” was followed from 24 hpf to 4dpf with images taken through the optical axis; and fish “C” was followed from 26 hpf to 4dpf with images taken through the equatorial axis of the lens.


Embryos were prepared for in vivo imaging as described previously for retinal imaging (Kay et al.,2004; Lohmann et al.,2005). Briefly, embryos were spawned and placed in Danieau's embryo medium (17.4 mM NaCl, 0.21 mM KCl, 0.12mM MgSO4, 0.18 mM Ca(NO3)2·4H2O, 1.5 mM HEPES, pH 7.6) at 22–28.5°C. Embryos were transferred to 0.2 mM 1-phenyl-2-thiourea (PTU) in Danieau's at 10–18 hpf to prevent melanophore development in the eye. Before imaging, embryos were anesthetized in 0.08 mM Tricaine, manually de-chorionated with forceps, and mounted in a Petri dish in 0.5% low melting point agarose dissolved in Danieau's, 0.2 mM PTU, and 0.08 mM Tricaine. Embryos were maintained in agarose with a layer of Danieau's, 0.2 mM PTU, and 0.08 mM Tricaine floating on top for up to 8 hr, or released from agarose and remounted for multiple imaging sessions.

Imaging was performed on an Olympus FluoView FV1000 multiphoton microscope with a long working distance Olympus 25× (1.05 NA) water immersion lens. Stacks of images through the lens were taken with Olympus Fluoview Ver1.7a software and saved as multi-tiff files. Lens equatorial width was measured with the Olympus Fluoview measurement tool. Image stacks were further analyzed using Amira 4.1.1 (Visage Imaging) including image segmentation, 3D projection views (“voltex”), and oblique re-slicing tools.


The rate of embryonic development was slowed as needed by keeping embryos at temperatures below 28.5°C (Kimmel et al.,1995). We have also found that different fish strains develop at different rates, for example, WIK embryos develop more slowly than the Q01 transgenic embryos. Embryonic ages in hpf listed in the manuscript are approximate and normalized to the previously published reports that the lens placode forms at 16 hpf and delamination occurs at 24–25 hpf (Soules and Link,2005; Dahm et al.,2007).

A note on terminology: because zebrafish eyes are oriented in a more lateral position than human eyes, it has been suggested that the usual description of “anterior” lens epithelium be replaced with the more accurate “distal” epithelium and that “posterior” pole be described as the “proximal” pole (Vihtelic,2008). We will continue to use the conventional anterior and posterior terminology for sake of clarity when comparing zebrafish and human lenses.


The authors appreciate the generous gifts of fish and advice from Dr. Rachel O.L. Wong and her laboratory. The authors thank David White for his technical expertise with zebrafish care and maintenance.