The eyes of fossorial mammals provide an excellent model for studies of eye malformations during development and for research into the evolution of blindness. The basic process of eye development is similar in all vertebrates and has been extensively reviewed (Graw,1996; Oliver and Gruss,1997; Jean et al.,1998; Chow and Lang,2001). In brief, the eye forms as a lateral outgrowth of the diencephalon (forebrain) neuroectoderm, which enlarges to become the optic vesicle. When the optic vesicle contacts the overlying surface ectoderm, an exchange of inductive signals between these tissues is thought to take place, which results in their coordinated invagination to form the lens vesicle and the optic cup (Chow and Lang,2001). The lens vesicle formed during this process is spherical and initially hollow, but soon becomes filled by the primary lens fibers, which form by elongation of the epithelial cells located at the posterior of the lens vesicle. Thereafter, new fibers are constantly added to the body of the lens from a population of proliferating cells located in the equatorial region. These new secondary lens fibers elongate and form the outer layers of the lens. As the lens grows, the primary fibers become compacted, and eventually become dissociated from the lens epithelium and the lens capsule and are pushed into the center of the lens, forming the embryonic lens nucleus. In the mouse, maturation of these primary fibers, which involves the loss of nuclei and organelles and synthesis of lens-specific proteins α-, β-, and γ-crystallins, is completed by postnatal day 14 (P14), when the eyelids open. The denucleation process is largely completed by P1, so that at birth, no nucleated fibers are found in the center of the lens (Vrensen et al.,1991). There is no turnover of proteins in the mature lens, but because the new fibers are added to it continuously, the expression of γ-crystallins can be detected in the maturing secondary fibers throughout the life of an animal. In the mouse, γ-crystallin expression increases until it reaches adulthood (about P40), and then gradually declines (Goring et al.,1992).
At the same time, and closely coordinated with the formation of the lens, morphogenesis of other eye structures takes place. The optic cup differentiates to form the neural retina and the retinal pigmented epithelium (RPE), and its anteriormost tips begin to form the iris and the ciliary body under the influence of the signals from the lens (Beebe,1986). The ciliary epithelium folds to form the ciliary processes, and the connective tissue of the ciliary body and the ciliary muscle are formed from the mesenchymal cells that have immigrated from the neural crest (Johnston et al.,1979). In the mouse, the first indication of ciliary folding and the appearance of the iris primordium are observed around E17, and by birth there are clear borders between the ciliary body and the retina posteriorly and between the ciliary body and the iris anteriorly (Pei and Rhodin,1970; Theiler,1989). The differentiation of neural retina begins around E17.5, with the appearance of the ganglion cell layer. By P7, the outer plexiform layer appears, separating inner and outer nuclear layers. The differentiation of the retina is completed by P14 (Pei and Rhodin,1970; Young,1984; Theiler,1989). Much of our current knowledge on eye morphogenesis is derived from the studies of mutant animals that exhibit various eye defects (Jean et al.,1998). Naturally blind animals can provide an additional model for such studies and allow researchers to circumvent the problems associated with inducing blindness in sighted animals. Since all naturally blind vertebrates must have evolved from sighted ancestors, investigations into their eye structure and development can also provide a better understanding of the evolutionary mechanisms operating during the atrophy of an organ that has become obsolete.
About 3.5% of all mammals are adapted to living underground and reduced visual systems are common among vertebrates adapted to such subterranean habitats (Nevo,1979). The loss of eyesight is postulated to be the result of the decrease in the evolutionary pressure that is responsible for the maintenance of functional organs of visual perception. Such a decrease in selection pressure would allow for the accumulation of mutations in genes involved in eye development (with the exception of developmental regulatory genes, which are also involved in the patterning and morphogenesis of other important structures in the body). As a result, various eye phenotypes can be observed among underground animals, ranging from relatively normal to completely malformed or microphthalmic. The smallest eyes, completely covered by skin, are seen in Notorychtes (marsupial moles), and in the golden moles Chrysochloris and Eremitalpa (Chrysochloridae; Insectivora). The eyes of Notorychtes are vestigial and lack lenses (Sweet,1906; Vaughan,1978), while those of Eremitalpa consist of a mass of disorganized cells representing the lens, surrounded by a well-differentiated retina (Gubbay,1956). The true moles of the family Talpidae have larger eyes with more typical eye architecture, including the iris, ciliary body, anterior and posterior chambers, and the retina with clearly distinguishable layers. The lens is present, but consists of irregularly shaped nucleated cells, which cannot be called true fibers (Slonaker,1902; Quilliam,1966). An interesting eye phenotype is observed in Spalax ehrenbergi (Spalacidae; Rodentia): both the pupil and the anterior chamber of this animal's eye are completely obliterated by an overgrowth of highly pigmented iris-ciliary body complex. The lens is very small, undifferentiated, and seems to undergo necrosis (Sanyal et al.,1990). These various eye phenotypes are thought to have appeared as independent evolutionary events, and their superficial similarity is postulated to be due to convergent evolution (Nevo,1979). For instance, members of the two rodent families Spalacidae and Bathyergidae have similarly reduced eyes, but they are more phylogenetically disparate than other families with sighted species within the order (Eisenberg,1981). While some of the structural and molecular aspects of the eye of the blind mole rat Spalax have been investigated (Quax-Jeuken et al.,1985; Hendriks et al.,1987; Avivi et al.,2001; Hough et al.,2002), very little is known about the Bathyergid eye. Previous descriptions of the structure of the adult naked mole rat eye are superficial and often contradictory (Cei,1946b; Hill et al.,1957). For instance, Cei (1946b) describes the mole rat lens as poorly differentiated with primitive characteristics, while Hill et al. (1957) state that the lens is differentiated; similarly, there is a disagreement between these two authors concerning the structure of the iris. To our knowledge, no description of the structure of the neonate mole rat eye or its development to adulthood has been published.
The naked mole rat (Heterocephalus glaber) is found in the hot, arid regions of Kenya, Somalia, and Ethiopia. Large colonies, usually composed of 75 to 80 related individuals, live in extensive burrow systems that can be up to 3 km long and occupy an area greater than 100,000 m2 (Sherman et al.,1991). Naked mole rats are eusocial, exhibiting a truly social structure with a reproductive division of labor, cooperative care of young, and an overlap of generations (Jarvis,1981). Since most naked mole rats within a colony do not reproduce, the definition of adulthood as reproductive maturity is difficult to apply. The youngest captive mole rats reported to be reproductively active were 8–12 months old (Jarvis,1991); therefore, for the purposes of this study, we have designated the adult as being over 12 months old, regardless of its reproductive status. The peculiar social system of the naked mole rats and their tendency to establish new colonies by fission result in extremely high levels of inbreeding both within a single colony and between the colonies of a particular geographic region. The inbreeding coefficient of these animals is the highest recorded among wild mammals and is similar to that for the inbred strains of laboratory mice (Reeve et al.,1990). This genetic homogeneity provides an additional advantage for the use of this animal in developmental studies, because there would be less developmental variation due to genetic background.
Here we provide a detailed description of the morphology of the adult naked mole rat eye and the developmental changes that accompany the growth and maturation of the ocular structures from birth into adulthood. We then examine the respective roles that cellular proliferation, programmed cell death, and fiber differentiation play in lens morphogenesis. Based on our results, a model for abnormal lens differentiation, and the role it plays in the morphogenesis of the eye in the naked mole rats, is proposed.
MATERIALS AND METHODS
Juvenile and adult naked mole rats were obtained from the Department of Zoology, University of Cape Town, where a number of successfully breeding colonies have been established (Jarvis,1981). Adult naked mole rats were sacrificed with chloroform or halothane and pups by decapitation. Eyes were harvested from mole rats at postnatal days 0, 1, 5, 14, 21, 32, and 34 and from 10 adults of various ages. All studies were carried out in compliance with the guidelines of the Animal Ethics Committee of the University of Cape Town.
Histology and Eye Measurements
The eyes of 19 mole rats, aged 1 day to 12 years, were analyzed using standard histological techniques. After dissection, the eye globe was placed in ice-cold fresh 4% paraformaldehyde (PFA) in phosphate-buffered saline. A sharpened tungsten needle was used to pierce the posterior hemisphere of the eyeball, allowing a rapid penetration of fixative into the eye. Eyes were fixed in PFA overnight, then dehydrated through a series of increasing ethanol concentrations and embedded in paraffin wax. The eyes were embedded so that sectioning would be through the vertical meridian of the eye. Sections 4 or 5 μm thick were stained with hematoxylin and eosin using standard procedures. Images were captured using a Carl Zeiss AxioCamHR digital camera mounted on an Axioskop 2 microscope. The measurements of the diameter of the eye, lens, ciliary body, iris, and cornea were made using the Axiovision 2.4 software package. To ensure consistency in the results, measurements of various structures within the eye were taken of histological sections through the center of the pupil. We recognize that histological processing causes tissue shrinkage, but since all our samples were processed in exactly the same way, we reason that this shrinkage would cause a consistent error in all measurements taken. The diameters of the whole eye and the lens were separately measured along the anteroposterior axis (at the level of ora serrata in the mole rat) and the dorsoventral axis (along the line extending from the cornea to the retina and passing through the center of the eye). The average values of these two measurements, for each eye and lens diameter, were then calculated. The lens flattening ratio was obtained by dividing the lens diameter, measured along the anteroposterior axis, by that along the dorsoventral axis. The lens flattening ratio reflects the degree to which the shape of the lens deviates from a perfect circle.
Scanning electron microscopy (SEM).
Whole eyes were fixed overnight in Karnovsky's fixative. The cornea, iris, and ciliary body were dissected out and returned to fixative for 1 hr. Samples were briefly washed in Sorenson's phosphate buffer, postfixed for 90 min in 1% osmium tetroxide, dehydrated, and dried at the critical point of carbon dioxide. Samples were splutter-coated with gold, mounted onto coated stubs, and examined with a LeoS440 scanning electron microscope operating at 15 kV.
Transmission electron microscopy (TEM).
Eyes were fixed overnight in Karnovsky's fixative and then cut in half equatorially and the anterior segment was returned to fixative for another hour. Samples were washed in Sorenson's phosphate buffer, postfixed for 90 min in 1% osmium tetroxide, stained for 30 min in 2% uranyl acetate, dehydrated, and embedded in epon-araldite or Spurr's resin. Sections 1 μm thick were stained for 5 sec with 1% toluidine blue in 1% borax. Ultrathin sections were stained in 8% saturated uranyl acetate and Reynold's lead citrate, then viewed and photographed using a JEM109 transmission electron microscope at 120 kV.
Apoptotic cells were identified using the in situ double-strand DNA break detection protocol developed by Gavrieli et al. (1992) with modifications. Wax sections, 6 μm thick, were placed on 3-aminopropyltriethoxysilane (APTES)-coated slides, dewaxed, and rehydrated according to standard procedures. The sections were then incubated in sodium sodium citrate (SSC, pH 7.0) for 20 min at 75°C, washed in distilled water (3 × 5 min), rinsed in terminal transferase buffer (1 × 5 min), and incubated with digoxygenin (DIG)-labeled dUTP and terminal transferase (Roche) for 2 hr at 37°C. Afterward, the sections were washed in distilled water (3 × 5 min), blocked in 5% sheep serum in tris-buffered saline (TBS) for 2 hr at room temperature, and incubated for 1 hr with an alkaline phosphatase-conjugated anti-DIG antibody (Roche). The antibody binding was visualized by the alkaline phosphatase reaction with 4-nitroblue tetrazolium chloride (NBT) and 5-bromo-4-chloro-3-indolyl phosphate (BCIP; all from Boehringer-Mannheim). The color reaction was allowed to develop for 1.5 hr and was subsequently stopped with Tris-EDTA.
Cell proliferation in eyes from animals of different ages was examined by phospho-histone H3 and bromodeoxyuridine (BrdU) analysis. For the phospho-H3 histone analysis, paraffin wax sections were dewaxed and hydrated according to standard procedures. For antigen retrieval, sections were microwaved for 1 min in 10 mM sodium citrate, allowed to cool for 1 min at room temperature, microwaved for a further 25 sec, and cooled for 20 min at room temperature. Sections were washed in buffer, permeabilized for 10 min in 0.2% Triton X-100 in PBS, blocked for 1 hr in 5% sheep serum at room temperature, and incubated in rabbit polyclonal antiphospho-histone H3 antibody (Upstate Biotechnology) overnight (1:500 in 5% sheep serum) at room temperature. Sections were washed in buffer, then incubated in secondary antibody (Cy-3-conjugated donkey antirabbit IgG in 5% sheep serum; 1:1,000; Amersham) in the dark overnight. All successive steps were conducted in the dark. Sections were counterstained for 10 min with 1:100 dilution of DAPI (Sigma) in PBS at room temperature, extensively washed in PBS, then mounted in Mowiol. Slides were stored at 4°C in the dark and viewed on a Zeiss Axioskop fluorescent microscope in the red and UV channels. Images were captured on a Carl Zeiss AxioCam camera mounted on an Axioskop 2 microscope using the Axiovision 3.1 software package.
For BrdU incorporation assays, we injected a 14-day-old mole rat with 300 μg/g body weight BrdU (Boehringer-Mannheim) for 2 days, and two adult mole rats, 4 and 12 years old, with 15 μg/g body weight at weekly intervals over a period of 14 weeks. All animals were sacrificed 24 hr after the last injection. Eyes were removed, fixed in 4% PFA, and processed to paraffin wax as described above. Paraffin sections 5 μm thick were dewaxed and rehydrated, washed in TBS, treated briefly with proteinase K (Roche), and denatured with 1.5 M HCl for 30 min at 37°C. Sections were blocked with 0.5% BSA in TBS for 30 min and incubated with a mouse monoclonal anti-BrdU antibody (Roche) diluted 1/50 in blocking solution overnight at 4°C. Sections were washed extensively in TBST and incubated with Alexa 488-conjugated goat antimouse secondary antibody (1/1,000 dilution in blocking solution; Molecular probes, Eugene, OR) for 1 hr in the dark at room temperature. Sections were washed extensively in buffer, counterstained for 10 min with DAPI (1:100 in TBS) at room temperature, then mounted in Mowiol. The slides were viewed and analyzed as described above.
For anti-γ-crystallin immunocytochemistry, 5 μm paraffin wax sections, cut through the center of the pupil, were placed on APTES-coated slides, dewaxed and rehydrated, rinsed with TBS, and incubated with 0.3% H2O2. Sections were blocked in 3% BSA in TBS for 1–1.5 hr at room temperature prior to incubation with 1/500 dilution of rabbit anti-γ-crystallin antiserum (generously provided by Dr. Linlin Ding, Joseph Horwitz Laboratory, Jules Stein Eye Institute, University of California, Los Angeles, CA) overnight at 4°C. The next day, the sections were washed extensively in TBST and incubated with peroxidase-conjugated swine antirabbit antiserum (Dako) for 2 hr. Antigen-antibody complexes were detected by 3,3′-diaminobenzidine tetrahydrochloride (DAB; Sigma) color reaction. Sections were lightly counterstained with hematoxylin, dehydrated through ethanol series, cleared in xylol, and mounted in Entellan (Merck).
Gel Electrophoresis and Western Blot Analysis
Lenses were dissected from four 2-day-old naked mole rat pups, homogenized in 20 μl extraction buffer (0.1 M Tris-HCl, 1% Nonidet P40, 0.01% SDS, 1 μg/ml aprotonin, and 0.1 mM phenylmethylsulfonyl fluoride) at pH 7.2 to extract the water-soluble proteins, and centrifuged at 12,000 g for 15 min. The water-insoluble proteins from the pellet were resuspended in 5 μl of 4 × sample dye (4% SDS, 20% glycerol, 10% β-mercaptoethanol, 0.125 M Tris, 0.03% bromophenol blue) and then diluted to 10 μl with extraction buffer. The supernatant (10 μl) was boiled to denature proteins, electrophoresed on a 12% sodium dodecyl sulfate polyacrylamide gel (SDS-PAGE), along with 5 and 1 μg samples of 2-day-old mouse lens protein extract, and transferred to a nitrocellulose membrane (Hybond-C, Amersham Life Science) at 8 V in 25 mM Tris-HCl and 5% methanol overnight. The membrane was blocked with 10% fat-free milk solution for 1.5 hr, and γ-crystallins were detected with 1/1,000 dilution of rabbit anti-γ-crystallin antiserum (provided by Dr. Linlin Ding). Primary antibodies were detected with horseradish-peroxidase-conjugated swine antirabbit polyclonal antibody using ECL fluorescent detection system (Amersham Life Science). After transfer, the gel was fixed in 20% methanol and 7% acetic acid solution overnight with constant agitation, washed three times in distilled water, and silver-stained for 15 min with 0.05 M solution of AgNO3. The protein bands were visualized after addition of the developer (0.005% citric acid, 0.02% formaldehyde). The reaction was stopped with 1% acetic acid solution, and the gel was photographed using ChemiImager.
Eye of Naked Mole Rat Is Microphthalmic
In this study, the structure and development of the naked mole rat eye from birth to adulthood were investigated in detail. The eyes from mice of corresponding ages were used for morphological and developmental comparisons. The mouse was chosen for this comparison because it belongs to the same order as the mole rat (Rodentia), has a comparable body size, good visual acuity, and well-studied eye structure. At birth, the mole rat greatly resembles the mouse in body size and general appearance (Fig. 1A). The eyelids of both species are closed at birth and first open 2 weeks later in the mouse (Graw,1996) and between 21 and 30 days after birth in the naked mole rat (O'Riain,1996). The eyes of the adult mole rat are deeply sunk into the head, and the eyelids are thickened and generally kept closed unless the animal is alarmed (Fig. 1B) (Jarvis,1991). The mole rat eye is typically round in shape, but lacks the turgidity characteristic of the mouse eye. This feature was particularly evident after dissection, and we noticed that the eyeball could be easily pinched or indented with forceps (much like squashing a football without air). The eye has an outwardly visible cornea and lens and a highly pigmented iris (Fig. 1C and D). The pupillary opening is somewhat irregular in shape. The posterior hemisphere of the adult mole rat eyeball, unlike that of the mouse, is always covered with a thick layer of fat (Fig. 1D). The normal extraocular muscles are present.
In comparison to the mouse, the adult mole rat has significantly smaller eyes (Fig. 1C). In order to compare the differences in size, we measured the diameter of both the unfixed whole eyes and the histological eye sections as described above. We found that, on average, the eye of the adult mole rat is two times smaller in diameter, and therefore eight times smaller in volume than the mouse eye (n = 9; Table 1). Thus, the naked mole rat eye can justifiably be termed microphthalmic.
Table 1. Comparative measurements of the body length and eye size of the mouse and the naked mole-rat
These values are calculated using the eye diameter measurements from histological sections.
Mouse neonate (n = 3)
Mole-rat neonate (n = 3)
Mouse adult (n = 4)
Mole-rat adult (n = 9)
We next examined the histological and ultrastructural features of the adult mole rat eye, using specimens from six mole rats aged 3 to 12 years, and four mole rats of unknown age. Samples of a variety of specimens of different ages are shown in Figure 2. The internal organization of the adult eyes is similar to that of the mouse, with the main structures—cornea, lens, retina, anterior chamber, iris, and ciliary body—being discernible. We did, however, find a number of significant differences related to the structure and cellular features of the cornea, iris, ciliary body, lens, and retina.
The mouse cornea is made up of a distinct corneal epithelium, stroma, and endothelium (Fig. 3A). The adult mole rat cornea has an epithelial layer, a proportionately thinner stromal layer, and, in normal histological sections, there appears to be no identifiable corneal endothelium present (Fig. 3B). To establish whether corneal endothelium is indeed absent, we carried out an ultrastructural examination using SEM and TEM. A corneal endothelial cell layer was clearly visible by SEM in a 5-year-old mole rat cornea (Fig. 4C). The cell morphology of this specimen appeared very different to that seen in the mouse. In the mouse, the corneal endothelium is a monolayer of tightly adhering hexagonal cells approximately 13.5 μm in diameter and forming a regular cobblestone pattern (Fig. 4A). In the 5-year-old mole rat, the endothelial cells are seen as a more rough layer, with the typical hexagonal pattern not evident (compare Fig. 4C to A). The cells are irregular and rounded in shape, with an average diameter of about 9.5 μm, and have many extending and overlapping processes. Many of the cells appear to have burst and secretory vacuoles are visible. The shape of these cells and their arrangement suggests that the cells are not joined together by junctional complexes, which are typically present in the mouse corneal endothelium (Kidson et al.,1999). Transmission electron microscopic examination of a cross-section through a 4-year-old mole rat cornea revealed the presence of large rounded vacuoles, enclosed between two neighboring cells, which were flattened and joined by tight junctions (Fig. 5A). Because such transcellular transport vacuoles and tight junctions are typical features of mouse corneal endothelium, we conclude that the true corneal endothelium cells are lining the anterior chamber of the naked mole rat eye.
When the mole rat eyes were bisected, we observed the lack of structured gel-like vitreous body. Instead, the eyeball was filled with a transparent liquid substance. Histological sections through the naked mole rat eyes showed that almost the entire vitreous chamber was filled by the extensively folded retina, which appeared to be closely associated with the posterior of the lens. The mole rat retina was folded to the extent never seen in the mouse, as if it was too big for the size of the eyeball (Fig. 2H). Because this condition of the retina was common in the adult mole rat, but was never seen in the mouse, we consider it a true difference between these species rather than a histological artifact. The general organization of the neural retina was similar to that of the mouse, which is composed of nine layers, namely, the photoreceptor layer, the external limiting membrane, the outer nuclear layer, the outer plexiform layer, the inner nuclear layer, the inner plexiform layer, the ganglion cell layer, the optic nerve fiber layer, and the internal limiting membrane (Fig. 3H and I). All except two of the retinal layers were well defined. The nerve fiber layer was absent in all adult mole rats examined and the ganglion cell layer appeared to be progressively reduced as the animals age (compare Fig. 3H to I). This reduction in the number of ganglion cells was apparent from examination of histological sections, and our observations have to be confirmed by ganglion cell counts. Besides the intensely staining ganglion cell bodies, the paler-staining large frothy cells, identified as microglial cells, were seen in increased numbers in the ganglion layer of the mole rat retina (Fig. 3I, arrowheads). The optic nerve contains cells arranged in an irregular pattern, which is different to the regular, more parallel arrangement seen in the mouse (Fig. 3D and E). Thus, while the retinal structure in the naked mole rat appears to be essentially normal, some histological features of reactive gliosis are evident.
One of the most prominent features of the mole rat eye is the large size and abnormal shape of the ciliary body. In the adult mouse eye, the ciliary body is folded into processes (Fig. 4D) and covered by two layers of columnar epithelium, the outer layer of which is pigmented and the inner unpigmented (Fig. 3O). The border between the ciliary body and the iris is clearly distinguishable. In the mole rat eye, however, typical ciliary processes were not observed, and there was no clear border between the ciliary body and the iris (Fig. 3P and Q). Overall, the ciliary body appeared flattened and elongated. In some specimens, extensive peripheral anterior synechias (abnormal attachment of the ciliary body to trabecular meshwork and to the peripheral posterior corneal surface) were observed (Figs. 2H and 3P and Q, arrows). Both the ciliary body and iris of all specimens were highly pigmented. The outer (pigmented) layer of the ciliary body shows slight folding, while the inner layer remains unfolded. The lack of ciliary processes was confirmed by our SEM studies. The adult mouse ciliary body (Fig. 4D) has macroscopic folds along its surface epithelium, whereas that of the 5-year-old mole rat is flat (Fig. 4E). Intriguingly, we observed deep pores or holes on the surface of the mole rat ciliary body, with an average width of 1.9 μm and length of 4.2 μm (Fig. 4F), and the average distance between these pore-like structures being 16.1 μm. The presence of these pore-like structures was confirmed by TEM, which showed intercellular spaces about 8 μm in diameter (data not shown).
The differences in the iris structure between the mole rat and the mouse became apparent during the examination of the freshly dissected eyes. Contrary to what is observed in the mouse, the shape of the pupillary border is irregular and not clearly defined; moreover, brown pigment granules are often observed in the cornea and within the anterior chamber, suggesting that adult mole rats exhibit iris degeneration and dispersion of pigment (Fig. 1C and D). These observations were confirmed by histological and ultrastructural analysis. In the mouse, the iris is composed of three layers, the anterior pigmented border, the stroma, and the inner pigmented epithelium. The iris muscle is clearly visible in the anteriormost tip of the iris (Fig. 3M, arrowhead). In some mole rats, the iris had the characteristic elongate shape and well-defined iris muscle, while in other specimens we observed the decrease in the thickness of the iris and degeneration of the pigmented epithelium (Fig. 3N, black arrowheads). The pigment, presumably released from these degenerating epithelial cells, was found in the anterior chamber angle and sometimes appeared to fill the trabecular meshwork (Fig. 2G, arrows). The degree of iris thinning varied greatly among the mole rats examined and did not appear to be directly related to their age.
The trabecular meshwork of some adults is very extensive, spanning the whole length of the elongated ciliary body. The trabecular beams of a 4-year-old specimen are pigmented and run almost parallel to each other, and large and well-developed aqueous channels are visible (Fig. 2G). In 8- to 12-year-old specimens, the size of the trabecular meshwork decreases dramatically, the trabecular beams are compacted and highly pigmented, and the spaces between them seem to be severely reduced (Figs. 2H and 3P). In some specimens, pigment-filled round structures were seen floating in the anterior chamber, or trapped in the trabecular meshwork (Figs. 2G and 3N, red arrows). A TEM micrograph of a conglomerate of these pigmented particles in a single cell, located within an intertrabecular space, is shown in Figure 5C. The finger-like projections on the surface of the plasma membrane of this cell, the shape of its nucleus, and the presence of a large number of lysosomes in the cytoplasm suggest that it is a macrophage.
The most striking difference between the mouse and the mole rat eye is the structure of the lens. The mouse lens is positioned in the anterior hemisphere of the eyeball and is held in place by the zonula fibers, which are also attached to the ciliary body. It is typically round in shape, with a characteristic pattern of nuclear distribution; a single layer of nucleated epithelial cells covers its anterior surface, and a group of nuclei belonging to the differentiating fiber cells is found at the equator (the bow region; Fig. 2F). The interior and posterior edge of the mouse lens are free of nuclei. In the adult mole rat, the lens appears to occupy most of the interior of the eyeball, and our observations during dissections and the absence of zonula fibers from all histological sections suggests that the lens floats freely inside the eyeball, rather than being attached to the ciliary body. In the majority of the specimens studied, the lenses were round, but often exhibited various irregularities in shape at the posterior margin or at the equator (Fig. 2G and H, arrowheads). A prominent lens capsule, up to 23 μm thick in older specimens, was present. The nuclei are found in a thin layer surrounding the entire lens. Our deduction is that it corresponds to the epithelial layer of the mouse lens, but it appears to cover a much more significant proportion of the lens surface. The bow region of fiber differentiation is often absent or very poorly developed in adult mole rats, and nuclei can be altogether absent from the region of the lens epithelium closest to cornea (Fig. 2H). When TEM was used to examine epithelial cell morphology, we observed that the cells appear to be very flattened and resemble squamous epithelium, rather than cuboidal epithelium that is found in adult mouse lens (Fig. 5E). The nuclei of these cells had very flat, pancake-like appearance. Since we did not observe an obvious bow region in the adult mole rat lens, BrdU incorporation assays were performed to establish whether new fibers would still be formed. Our results (Fig. 7C and D) indicate that some of the cells located within the posterior third of the epithelium were dividing, and that the mole rat epithelium was thus still mitotically active even in 12-year-old mole rats. TEM of the lens equator in a 4-year-old mole rat confirmed the presence of maturing lens fibers, which still contained some of the organelles and exhibited highly variable thickness and length (Fig. 5B). Therefore, despite its abnormal architecture, the naked mole rat lens appears to maintain its growth and fiber differentiation activity throughout the life of the animal.
Postnatal Development of Naked Mole Rat Eye
In order to begin to define the sequence of the major developmental events in the postnatal mole rat eye, we examined the structure and histology of the naked mole rats of different ages, ranging from neonate to 34 days old. Due to the peculiar social structure of the naked mole rats, breeding females are essential for the survival of the colony and consequently could not be sacrificed to obtain embryos; therefore, with the exception of one batch of embryos at a very late developmental stage, the neonate was the earliest developmental stage available for our investigations.
In newborn mice, the anterior chamber is fully formed, and the squamous endothelium is clearly visible on the inner surface of the cornea (Fig. 2A). In the newborn mole rat eye, however, there was no space separating the cornea from the lens. In fact, in all three neonate specimens examined, the cornea appeared to be attached to the anterior surface of the lens and the ciliary body (Fig. 2B). In order to establish when the anterior chamber is first formed, histological sections from juvenile mole rats of various ages were examined. The presence of a space between the cornea and the lens was noted in a 14-day- and three 32- to 34-day-old mole rats, while no such space was observed in a 21-day-old mole rat (Fig. 2C–E). Therefore, we deduced that the time of anterior chamber formation is variable, and that the process is usually completed by the time mole rats are 32 days of age. It has been proposed that the formation of the anterior chamber in the mouse is concurrent with the corneal endothelium formation (Kidson et al.,1999; Reneker et al.,2000). In an attempt to determine whether these processes also occur concurrently in the mole rat, we examined semithin (0.5 μm thick) resin sections of a late embryonic, 5-, 12-, 14-, and two 34-day-old mole rats eyes. Unfortunately, we were not able to observe anterior chamber in any of these sections because the lens was always found to have changed its position considerably with respect to other ocular structures, sometimes being turned so that its anterior epithelium was facing the retina. Thus, the space between the lens and the cornea in these specimens appears to be a resin-embedding artifact (except in the embryo, where the lens is not displaced and is still adherent to the cornea). Since we never observed such lens displacement when mouse eyes were processed in an identical way, it appears that the mole rat lens is not firmly attached to the cornea even in those specimens that apparently lack anterior chamber. The typical flattened endothelial cells filled with transcellular vacuoles were observed in both the central and the peripheral regions of the cornea in the 34-day-old and the 14-day-old specimen. In the 12-day-old mole rat, the flattened endothelial cell layer was observed, but transcellular vacuoles were not found. In the 5- day-old specimen, corneal endothelial cells with transcellular vacuoles were seen in the peripheral regions of the cornea, but not in the center. No endothelial layer was seen in the embryonic specimen, and the space between the lens and the cornea contained stromal cells, which appeared to adhere to the lens. TEM and SEM was used to confirm these observations. True corneal endothelial cells were seen in the peripheral regions of the TEM sections of a 5-day-old specimen, and in the cornea of a 32-day-old specimen (data not shown). A corneal endothelial cell layer with a cell morphology closely resembling that of an adult mole rat could also be seen in the scanning electron micrograph of a 21-day-old mole rat cornea (Fig. 4B). These findings suggest that, contrary to what is observed in the mouse, the formation of the anterior chamber in the naked mole rat occurs after the formation of the corneal endothelium.
In the mouse, the ciliary body and iris morphogenesis starts prenatally, at about E17 (Theiler,1989), and continues until P10, when both of these structures are morphologically mature (Monaghan et al.,1991; Smith et al.,2001). The ciliary body and iris primordia are distinguishable by E18, and there is a clear boundary between these structures in the neonate mouse (Fig. 3J). Contrary to what is observed in the mouse, there are no clearly distinguishable ciliary body and iris primordia in the eye of the neonate naked mole rat, though a highly pigmented structure covered with a nonpigmented epithelial layer is clearly visible (Fig. 3K) and probably represents the common primordium of the iris and ciliary body. This ciliary body-iris complex of the mole rat is greatly enlarged, making up about 30% of the eye circumference, while in the mouse the combined lengths of the ciliary body and the iris make up only 8% of the circumference (Table 2). By the time the naked mole rat is 14 days old, the ciliary body appears to have become distinct from the iris (Fig. 3L). Interestingly, the combined length of the iris and the ciliary body at that age is approximately equal to the length of the neonate ciliary-body-like complex (Table 2), suggesting that in the mole rat, part of this initial structure differentiates to form the iris.
Table 2. Comparative measurements of the eye structures
Eye diameter (mm)
Lens diameter (mm)
Lens flattening ratio
Corneal thickness (μm)
Iris length (mm)
Ciliary body length (mm)
Mole-rat neonate (n = 3)
Mouse neonate (n = 3)
14-day old mole-rat (n = 1)
21-day old mole-rat (n = 2)
32-day old mole-rat (n = 1)
Adult mouse (n = 4)
Adult mole-rat (n = 9)
Ciliary fold morphogenesis in the mouse begins around E18, and the elongated adult-like ciliary processes are evident by P4 (Theiler,1989; Smith et al.,2001). In the neonate mole rat, there is no evidence of the ciliary fold formation: only the outer (pigmented) layer shows some folding, while the inner (nonpigmented) layer remains flat and accommodates the folds of the pigmented layer by the decrease in the height of the cells lying directly above the folds (compare Fig. 3J and K). True folding of the ciliary epithelium, similar to that of the neonate mouse, is observed in the 14-day-old mole rat (Fig. 3L). Our BrdU-labeling experiments show large numbers of proliferating cells in the inner (pigmented), but not in the outer (nonpigmented) epithelial layer of the ciliary body in the 14-day-old mole rat (Fig. 7A and B, arrowheads), suggesting that increased levels of cellular proliferation in the inner epithelial layer can be responsible for the ciliary folding. However, as the animal grows, the ciliary processes appear to become flattened out. The absence of the typical folding of the ciliary body epithelium into processes in the 21-day-old mole rat was confirmed by scanning electron microscopy. The surface of the 21-day-old mole rat ciliary body is flat and very similar to what is observed in the 5-year-old mole rat (data not shown).
In the mouse, formation of the trabecular meshwork is a postnatal event, which starts around P6 and reaches its full structural and functional maturity by P21–P42 (Smith et al.,2001). The mole rat trabecular meshwork appears to form at about the same time as the anterior chamber, when the ciliary body separates from the cornea. Well-defined trabecular beams and aqueous channels were visible in the eye angle of a 32-day-old mole rat (Fig. 2E), but in younger specimens the state of differentiation of this structure cannot be easily observed due to the adherence of the iris and ciliary body to the cornea and consequent absence of the iridocorneal angle. The trabecular meshwork in all juvenile mole rats examined remained unblocked, and no signs of pigment dispersion are visible, suggesting that this phenotype develops slowly as the animal ages.
While the anterior chamber development is apparently delayed in the mole rat when compared to the mouse, the reverse is true for the retina. The retina of the neonate mouse is not yet fully differentiated, and only RPE, outer neuroblastic layer, inner plexiform layer, ganglion cell layer, and forming nerve fiber layer can be clearly distinguished (Fig. 3F). In the neonate mole rat, on the other hand, all 10 retinal layers characteristic of the adult are fully formed (Fig. 3G). Traces of nerve fiber layer, which are not distinguishable in the adult mole rat, can be seen in the neonate mole rat retina (Fig. 3G, arrowhead).
The neonate mole rat lens is oval-shaped, which is similar to what is observed in the neonate mouse (Table 2, Fig. 3R and S). However, as the mole rat grows, the shape of the lens becomes more irregular, with an abnormally defined, collapsing posterior margin observed in 21- and 32-day-old mole rat lenses (Fig. 2D and E). Although the shape of the lens and the degree of collapse vary greatly among specimens, it was consistently observed in the juvenile mole rat specimens (n = 5). High-magnification examination of these abnormal lenses revealed that the individual fibers at the posterior margin were bent or misshapen and were not arranged in the regular fashion characteristic of the mouse lenses. We also noticed significant differences in the nuclei distribution in the lenses of these two species. In the mouse, a layer of nucleated epithelial cells covers the anterior surface of the lens, and there are accumulations of nuclei at the bow region, located at the equator of the lens (Fig. 3S, boxes). In the neonate mole rat, the epithelial cell layer appears to extend over a significantly larger area of the lens. The region of fiber differentiation (the bow region), identified by the elongated shape and the clustered arrangement of their nuclei, is shifted toward the posterior pole of the lens (Fig. 3R, boxed regions). Moreover, in the mole rat, nucleated cells are also found throughout the center and the posterior of the lens (Fig. 3R, arrowheads), while in the mouse these areas contain only mature fibers and are therefore nuclei-free (Fig. 3S). Interestingly, in juvenile mole rats, the center of the lens gradually becomes nuclei-free, suggesting that the fiber maturation process is delayed in this animal compared to the mouse. The lens epithelium in both the neonate mouse and the mole rat is a simple cuboidal epithelium. However, in the mole rat, the morphology of the lens epithelial cells changes from regular cuboidal, with round nuclei in a 5-day-old specimen (Fig. 5D), to pseudostratified cuboidal in a 32-day-old (data not shown), until the cells become completely flattened and resemble squamous epithelium in the adult (Fig. 5E).
Molecular Basis of Abnormalities in Mole Rat Lens
To begin to investigate the molecular basis of the structural abnormalities in the naked mole rat lenses, we carried out a series of experiments designed to establish whether these abnormalities are due to changes in cell death, proliferation, or differentiation.
Using the TUNEL reaction to identify apoptotic cells, we found no evidence of cells undergoing programmed cell death in the lens in any of the neonate eye sections examined (Fig. 6A). However, in a 21-day-old mole rat, a number of nuclei at the posterior margin of the lens were TUNEL-positive (Fig. 6B, arrowheads). The staining was considerably fainter than in our DNase-treated positive controls (data not shown). The nuclei of differentiating lens fibers undergo DNA fragmentation and therefore can be labeled by TUNEL technique under certain experimental conditions (Bassnett and Mataic,1997; Ishizaki et al.,1998; Wride and Sanders,1998), though the staining is not as intense as that of apoptotic cells. Moreover, darker staining apoptotic cells were observed in the ganglion cell layer in the neonates and in the inner nuclear layer of the 21-day-old mole rats (data not shown). We conclude that increased levels of apoptosis are not observed in the naked mole rat eye.
We next carried out labeling studies (phospho-histone H3 and BrdU incorporation) to obtain a picture of the zones of cell proliferation in the lens. In particular, we were interested in establishing whether the nucleated cells located at the posterior margin of the lens were proliferative cells. Eyes from three juvenile mole rats were examined (two 32-day-old mole rats and one 14-day-old mole rat). In all of the mole rats examined, we found evidence of dividing cells in the region of the epithelium located just posteriorly to the lens equator (Fig. 3A). None of the nuclei in the center or posterior border of the lens were BrdU-positive. These results suggest that the overall pattern of proliferation is similar in the mole rat and in the mouse, but the position of the proliferative compartment (equatorial or bow region) in the mole rat is shifted toward the posterior end of the lens. These results led us to the next question of whether the presence of nuclei in the center of the mole rat lens, as well as the abnormal shape of the lens fibers, can be attributed to a defect in lens fiber differentiation.
γ-crystallin synthesis in mammals is the essential part of the lens fiber differentiation process. In the mouse, the synthesis of γ-crystallins commences at about 14 day of embryonic development and continues until the mouse reaches reproductive age (about P40), after which it gradually declines (Goring et al.,1992). The proteins persist in the lens fibers throughout the life of the animal, and their presence in the lens at appropriate concentrations is essential for maintaining its transparency. In order to investigate further the process of lens differentiation in the naked mole rats, we used SDS-PAGE and Western blotting to analyze the 2-day-old and the 6-year-old mole rat lens protein extracts for the presence of γ-crystallins. The adult (6-month-old) and 2-day-old mouse lens extracts were used as positive controls. A protein band of the expected molecular weight of 21–22 kDa, corresponding to all six γ-crystallin proteins (Siezen et al.,1988), was observed in the immunoblot of the mouse lens extracts (Fig. 8B, lanes 7 and 8). Our results show that at least some of the γ-crystallins are present in the soluble protein fraction from the lenses of a 2-day-old mole rat (Fig. 8B, lane 4), but we were unable to detect crystallins in the lens of a 6-year-old mole rat (Fig. 8B, lanes 1 and 2) despite repeated analysis. Moreover, the diffuse appearance of the 6-year-old mole rat protein bands on the gel is an indication of increased protein degradation.
The distribution pattern of γ-crystallins within the lenses of mole rats of different ages was analyzed by immunocytochemistry (ICC), and the results are shown in Figure 9. Sections of adult and neonate mouse eyes were used as positive controls. In the mouse, γ-crystallins are expressed throughout the lens, with the exception of the epithelium and the equatorial region (which includes fibers that have just begun differentiating; Fig. 9F and G). The pattern of γ-crystallin expression in the mole rat was found to be different to that of the mouse. In the neonate mole rat, most of the lens fibers stain positive for γ-crystallins, leaving the epithelium and a thin circle of two to three fiber layers extending around the exterior of the lens (presumably corresponding to the new differentiating fibers) signal-free (Fig. 9B). In the 21-day-old mole rat, this γ-crystallin-free zone was extended to 7–10 layers (Fig. 9C), suggesting that fibers formed, after the mole rat was born, do not express this protein. Only the center of the lens of a 3-year-old mole rat stained positive for γ-crystallin (Fig. 9D), while no staining was observed in the lens of the 12-year-old adult (Fig. 9E). We conclude that the synthesis of γ-crystallins in the mole rat lens is terminated around the time of birth, and that the γ-crystallin produced earlier undergoes degradation as the animal ages. This process could be responsible for the irregular shape and fiber arrangement observed in the mole rat lenses.
Behavioral observations on the naked mole rat suggest that this animal relies almost solely on olfactory and tactile cues to navigate and when interacting with conspecifics. There is no evidence that they rely on visual information for any aspect of their daily lives (Narins et al.,1997). Our histological examination of the naked mole rat eye lends support to these observations. The small eye size of these animals probably results in their having a very restricted visual field. In addition, the irregular shape of the lens, and the presence of cellular nuclei along the visual axis, might suggest that light scattering prevents clear images from forming on the retina. The absence of the zonula fibers and the reduced ciliary muscle of most mole rat specimens suggest that light focusing on the retina does not occur. However, the presence of rudimentary iris muscle indicates that the mole rats are able, at least to a degree, to regulate the amount of light that enters the eye. Thus, we conclude that the visual abilities of the naked mole rats are limited to judging the intensity of the surrounding light (i.e., being able to distinguish between night and day, or between being inside the burrow or outside it) and, possibly, to seeing the shadows cast by large moving objects, without being able to see their details.
The above findings lead us to the question of why, after at least 25 million years of subterranean evolution (Bennett and Faulkes,2000), the naked mole rats still retain all of their ocular structures and apparently a degree of visual ability? The conservation of the eye architecture of this species is especially surprising when compared to the regressed ocular phenotypes seen in other dark-adapted vertebrates. For instance, the eyes of the blind cavefish Astyanax mexicanus do not develop a cornea, an iris, secondary lens fibers, or differentiated retina (Yamamoto and Jeffery,2000; Jeffery,2001). The cornea, iris and ciliary body epithelia, differentiated retinal layers, as well as the vitreous and aqueous chambers are altogether absent from the eyes of the marsupial mole Notoryctes typhlops (Sweet,1906) and the insectivorous moles Scalops aquaticus and Eremitalpa granti (Slonaker,1902; Gubbay,1956). These extremely reduced eyes are thought to have evolved in response to the evolutionary pressure to decrease the metabolic expenditure, associated with the formation and maintenance of the organ that is no longer used (Nevo,1998). However, the naked mole rats, as well as a number of other fossorial mammals, retain much of the normal ocular architecture and, in particular, an apparently normal retina. This suggests that retaining the capacity for light-dark discrimination is important for the survival of these animals. The soil-removal activity of the naked mole rats results in their direct exposure to sunlight, as the animals kick soil out of an open mound. The open mound poses a further threat of exposure to aboveground predators (Sherman et al.,1991). An ability to detect light and dark and sudden transitions associated with the arrival of a predator at well-lit burrow entrance may confer a survival advantage and hence be maintained by natural selection.
It is interesting to notice that the degree of eye reduction in various burrowing species correlates very well with the method they use for soil digging. Thus, the eyes of the animals that use their head to push the excavated soil (Spalax), or to force their body forward through the soil loosened by their forelimbs (Notoryctes, insectivorous moles), are much more reduced than those of the animals that use both their incisors and forelimbs to scrape off the substrate and push the loosened soil backward with their limbs (Bathyergidae, Rhyzomyidae) (Vaughan,1978; Nevo,1979; Webb et al.,1979; Bennett and Faulkes,2000). The forelimb diggers appear to be more at risk of eye damage and infection than the teeth diggers. It is possible that one of the major selective forces favoring the reduction in the eye structures in subterranean mammals is the need to protect this soft and sensitive organ from the abrasive effects of soil.
Studies to date on chick (Hay and Revel,1969) and mouse eyes (Kidson et al.,1999) suggest that the formation of the anterior chamber is coupled to the morphogenesis of the corneal endothelium. Thus, the anterior chamber forms as the neural crest cells, closest to the lens, undergo mesenchymal-epithelial transformation and become the tightly packed corneal endothelial layer, which separates the extracellular matrix of the corneal stroma from the surfaces of the lens and iris (Kidson et al.,1999; Reneker et al.,2000). Interestingly, however, in the naked mole rats, the anterior chamber appears to form significantly later than the corneal endothelium. Thus, it seems that the formation of the corneal endothelial layer is insufficient for the establishment of the proper anterior chamber architecture.
The anterior chamber of the adult mouse is filled with aqueous fluid, which is produced by the ciliary body. However, when the anterior chamber is first established (E15 in the mouse), the ciliary body has not yet formed (Pei and Rhodin,1970; Theiler,1989). Therefore, it is currently not clear what the origin of the fluid that fills the anterior chamber is when it first appears. Perhaps the reason for the delayed formation of the anterior chamber in the naked mole rats is the fact that there is nothing to fill the anterior chamber until the ciliary body is mature enough to synthesize the aqueous humor. The molecular mechanisms responsible for this phenotype are still to be elucidated. The close similarity of the mole rat anterior chamber architecture (adherence of the base of the ciliary body/iris to the cornea, low amplitude of the anterior chamber, poorly developed iridocorneal angle) to the phenotype of the Lmx1b−/− mouse mutants (Pressman et al.,2000) suggests that altered Lmx1b expression could play a role. Interestingly, the lack of or decrease in the size of the anterior chamber occurs among other fossorial mammals with reduced eyes (Rhizomyidae) (Cei,1946a), suggesting that a similar change in the eye development could be responsible for the convergent evolution in these subterranean mammals.
Both the iris and the ciliary body of the naked mole rats are highly pigmented, and the ciliary body is very large relative to the size of the eye. The extreme pigmentation and enlargement of these anterior chamber structures appears to be very common in the mammals adapted to a fossorial lifestyle. This phenomenon is most noted in Spalax ehrenbergi and Notoryctes typhlops, where the pupil is completely obliterated by a mass of pigmented tissue (Sweet,1906; Cei,1946a; Quilliam,1966; Sanyal et al.,1990). The presence of enlarged ciliary body could be of adaptive value to these animals, as it probably protects the retina from sudden exposure to bright light when the animal emerges from its burrow during the day (Sanyal et al.,1990). Alternatively, the increase in the relative ciliary body size might not in itself confer any evolutionary advantage, but result from the selective pressure to decrease the size of the retina. As have been pointed out previously, neuronal tissue uses larger amounts of energy than most other tissues, and therefore maintenance of excess neurons is an evolutionary luxury that is strongly selected against (Nevo,1998).
In the mouse, the iris and ciliary body epithelia are formed from the tip of the optic cup, while the iris stroma and the ciliary muscle are derived from the cephalic neural crest (Beebe,1986). A part of the anterior optic cup is specified as the ciliary epithelium at around E12 by yet unidentified signals from the lens epithelium (Genis-Galvez,1966; Stroeva,1967; Beebe,1986; Thut et al.,2001). The morphologically distinct ciliary body is first formed in the mouse around E16.5–17, and the formation of the ciliary processes and its functional maturation is completed postnatally (Theiler,1989). In the neonate mole rat, the size of the ciliary body and iris, relative to the eye circumference, is three times greater than in the newborn mouse (Table 2). This suggests that, in the naked mole rat, a greater proportion of the anterior optic cup is instructed to adopt the ciliary body/iris fate during embryogenesis. It is interesting that the area of the lens epithelium, the proposed source of the ciliary body-specifying signal, is also increased in the naked mole rat. Possibly, this extended area of lens epithelium is responsible for the establishment of the larger ciliary body.
As in the mouse, the ciliary body and iris of the naked mole rats are not mature at birth. In fact, there is no distinguishable border between the ciliary body and the iris in the neonate mole rat. Our measurements (Table 2), however, suggest that the iris primordium is present in the neonate mole rats, because the length of the ciliary body, as measured on the juvenile and adult mole rats, is 0.4–0.7 times that of the initial ciliary body-like structure of the neonates (Table 2). This piece of data suggests that the iris specification in the mole rat also occurs during the embryonic life, similar to the mouse, even if the morphology of this structure is initially quite different. As in the mouse, most of the growth of the iris in the mole rats occurs postnatally. We observed thinning of the iris stroma and loss of the iris pigment in many old mole rat specimens. This interesting phenotype could be a consequence of the elevated levels of melanogenesis in the iris. The increased eye pigmentation in the naked mole rats suggests that the melanocytes of the anterior surface of the iris, which are not active in the adult mouse, are still producing pigment in the adult mole rats. These greatly increased levels of pigment synthesis could lead to the melanocytes becoming filled up with pigment, bursting and releasing their contents into the anterior chamber. This is very similar to some of the phenotypic features of the mouse models of pigment-dispersion glaucoma (John et al.,1998; Anderson et al.,2002). Genetic dissection of the mechanisms regulating the pigment synthesis in the naked mole rats could provide additional insights into the etiology of human pigmentary glaucoma.
The ciliary body of all adult and most juvenile mole rat specimens is flat and elongated in shape, and the ciliary processes are absent. This lack of the ciliary processes is another feature that mole rats share with other subterranean mammals [e.g., the common mole (Quilliam,1966)], but not with any other vertebrate models exhibiting normal visual acuity, i.e., chick, mouse, frog Rana (Beebe,1986). It is currently not clear whether this lack of the ciliary processes affects the functioning of the ciliary body in a way that is somehow advantageous to these animals, or whether it is simply a developmental consequence of the altered signaling within the eye, leading to the increased ciliary body size. The ciliary fold morphogenesis in chick appears to be dependent on the intraocular pressure (Bard and Ross,1982). Our observations during dissections of the mole rat eyes suggest that its intraocular pressure is reduced, which could be another factor contributing to the absence of the ciliary processes.
The mouse trabecular meshwork is derived from a group of mesenchymal cells situated in the angle between the base of the iris and the cornea. These cells undergo differentiation to form trabecular beams separated by channels, which serve to allow the exit of the aqueous fluid from the anterior chamber. The process of trabecular beam differentiation commences around P10 in the mouse, and the mature meshwork is established by P21 (Smith et al.,2001). In the mole rat, the formation of the trabecular meshwork cannot be assessed at early postnatal stages due to the absence of the anterior chamber and the consequent adherence of the iris and ciliary body to the inside of the cornea, obliterating the iridocorneal angle. It appears, however, that in the 5-day-old specimen, no identifiable trabecular beams and channels are yet evident, making the state of differentiation of that specimen comparable to P6–P10 mouse. At 30–34 days of age, the trabecular meshwork (TM) of the mole rat is extensive and well-differentiated and comparable to the mouse TM at P21–30. Moreover, no TUNEL-positive cells were noticed in the TM of the 21-day-old specimen, suggesting that, similar to the mouse but contrary to what has been reported in the rats, the TM development in the mole rats does not involve apoptosis (Smith et al.,2001). A very unusual feature of the naked mole rat eye, often observed in mature specimens, is a closure of the eye angle and degeneration of the trabecular meshwork. These changes could be related to the unusual longevity of the mole rats (Buffenstein and Jarvis,2002), but the extensive tissue loss from the trabecular meshwork observed in these animals is not a feature of the normal aging process even in humans (Oates and Belcher,1994). It is possible that increased macrophage activity associated with the pigment dispersion results in the age-related deterioration of the trabecular meshwork.
The structure and retinal layer organization in the mole rat eyes are essentially normal, except that, in the adults, no nerve fiber layer and an apparent decrease in the density of the retinal ganglion layer are observed. Retinal ganglion cells appear to be the most susceptible to apoptosis as the result of increased intraocular pressure, as seen in glaucomas (John et al.,1998). It is possible that, in the mole rats, the impaired circulation of the aqueous fluid due to the blockage and degeneration of trabecular meshwork causes the loss of retinal ganglion cells. However, the reduction in the numbers of ganglion cell layers has been observed in other completely or partially subterranean mammals (Herbin et al.,1994) and is not accompanied by the pigment-dispersion phenotype seen in the naked mole rats. Other factors, perhaps related to decreased exposure to light in subterranean habitats, could therefore play a role. In order to investigate the age-related changes in retinal architecture, ganglion cell counts and extensive quantitative studies of apoptosis in the retina of mole rats of different ages will need to be performed.
Cellular Organization of Mole Rat Lens Is Abnormal
Our histological studies demonstrated that the nuclear distribution in the naked mole rat lens differs significantly from that of the mouse lens. In the neonate and juvenile mole rats, the nuclei are found in a thin layer covering about two-thirds of the anterior surface of the lens (presumably corresponding to the nuclei of the lens epithelium in the mouse), as well as within the posterior hemisphere of the lens, which in the mouse is occupied only by nonnucleated lens fiber cells. In order to ascertain whether this peculiar nuclear distribution is the result of the failure of the prospective lens fiber cells to exit cell cycle, we assayed the levels of cellular proliferation using BrdU incorporation. Because continuous proliferation of lens fiber cells resulting from failure to exit cell cycle activates an apoptotic response (Morgenbesser et al.,1994; Lahoz et al.,1999), we performed TUNEL assays to determine the levels of programmed cell death in the mole rat lens. We did not find any BrdU-positive or TUNEL-positive cells within the fiber cell compartment. This suggest that the abnormal presence of the nucleated cells within the center of the lens is due to the delayed nuclei degradation in the maturing fibers, rather then to the continuous cell proliferation and failure to exit the cell cycle. The only proliferating cells that were found in the mole rat lenses were localized at the equator. This suggests that the nuclei belong to the epithelial cells of the lens proliferative region, which appears to be displaced posteriorly in the mole rat compared to the mouse.
In the mouse, the lens polarity is established early in the development, around E12, when the posterior cells of the lens vesicle are instructed to elongate and form lens fibers, while the anterior cells remain epithelial. The region of high cell proliferation is established just anterior of the equator, and the cells, which are produced there throughout the life of the animal, move posteriorly and undergo differentiation to form the secondary lens fibers (Pei and Rhodin,1970; Theiler,1989; Graw,1996). Classic transplantation experiments have demonstrated that molecular signals originating from the retina and present in the vitreous humor of the eye are responsible for the establishment of the lens polarity (Coulombre and Coulombre,1964; Yamamoto,1976). More recent tissue culture-based studies identified that a high concentration of a group of diffusible signaling molecules—fibroblast growth factors (FGFs)—is able to cause lens fiber differentiation, and a lower concentration can initiate proliferation of lens epithelial cells (McAvoy and Chamberlain,1989). The existence of the FGF concentration gradient in vivo, compatible with its role as the inducer of cell proliferation and fiber differentiation in the lens epithelium, was demonstrated within the ocular media. Moreover, transgenic mice-based studies indicate that increase in the FGF concentration within the eye leads to differentiation of lens fiber cells in the anterior region normally occupied by the lens epithelium (Chamberlain and McAvoy,1997; Lang,1999). Since the role of FGFs in lens morphogenesis has been documented in a range of mammalian species (human, rat, mouse, cow) (Chamberlain and McAvoy,1997), it is likely that a similar mechanism operates in the mole rats. We suggest that the displacement of the proliferative region toward a more posterior location in the naked mole rats could be caused by the decreased FGF concentration in the vitreous humor. Further investigations are necessary in order to ascertain what role FGFs and their receptors play in the establishment of the lens polarity in this animal.
The unusual persistence of nuclei in the central fibers of the mole rat lens led us to ask whether other aspects of lens fiber differentiation are abnormal. We therefore used ICC and Western blots to investigate the expression of the lens-specific proteins γ-crystallins, which are synthesized only in lens fiber cells and thus can be used as a marker of lens differentiation. In our Western blot experiments, we detected γ-crystallins in the lenses of 2-day-old mole rat pups, but not in the lens of the adult (6-year-old) mole rat. This suggests that either the synthesis of this protein is turned off in the adult lens, or the protein undergoes degradation. This result was confirmed by our ICC experiments, which showed that at birth, γ-crystallins were expressed throughout the entire fiber compartment of the lens, with the exception of a few lateral fibers, which appear to be still undergoing the differentiation process. The same pattern of the γ-crystallin staining is observed in the mouse eyes. However, as the mole rats matured, the area of the γ-crystallin-free fibers was expanded. Moreover, only a very weak signal was detected in the center of the 3-year-old mole rat lens, and no γ-crystallin-positive fibers were observed in a 12-year-old mole rat. It therefore appears that the synthesis of γ-crystallins is downregulated soon after birth, and also that the protein that was already synthesized becomes degraded as the animals age. Since we did not investigate the levels of γ-crystallin mRNA, we could not establish whether the turning-off of γ-crystallin synthesis is due to a translational or a transcriptional downregulation event. Six individual γ-crystallin proteins are synthesized in the mouse and rat lenses (van Leen et al.,1987; Siezen et al.,1988; Goring et al.,1992), and promoter and knockout studies indicate that all γ-crystallins are under common transcriptional control by a number of factors, including c-Maf (Kawauchi et al.,1999; Kim et al.,1999; Ring et al.,2000) and Sox1 (Nishiguchi et al.,1998). Therefore, it is possible that a mutation in one of these regulators can result in the observed absence of γ-crystallin synthesis.
The expression pattern of the γ-crystallins in the naked mole rat is unusual even when compared to that of other blind subterranean mammals. For example, the common mole, Talpa europaea, despite having a small poorly differentiated lens, consisting of what appears to be nucleated primary fibers, shows a typical lens expression of α-, β-, and γ-crystallins (Quax-Jeuken et al.,1985). The undifferentiated lens of the blind mole rat, Spalax ehrenbergi, does not express either β- or γ-crystallins, but does exhibit α-crystallin expression (Quax-Jeuken et al.,1985; Hough et al.,2002). In this animal, the lens development appears to be arrested earlier than in the mole, so that no fiber differentiation (and thus no synthesis of fiber-specific proteins) takes place. It is rather difficult to speculate why there is a switching-off of γ-crystallin expression in the naked mole rat, unless we assume that formation of a lens that allows at least some of the light through gives an evolutionary advantage to the naked mole rat. γ-crystallins differ from all other crystallin types by having a relatively high phase-separation temperature, and thus a tendency to precipitate when the body temperature is slightly lowered, as observed in cold cataract formation in neonate rats and mice (Siezen et al.,1988). The naked mole rats, unlike any other mammals, are poikilothermic, i.e., have a variable body temperature (Buffenstein and Yahav,1992). It is possible, therefore, that switching off the expression of γ-crystallins was an evolutionary solution to producing a lens that would not become completely opaque when the body temperature of the mole rat fluctuates. Sequencing mole rat γ-crystallin genes to determine the cryostability of the proteins they encode, and investigating γ-crystallin expression in warm-blooded members of the family Bathyergidae, could help clarify this matter.
As can be clearly seen on histological and even on ICC sections, the lens fiber arrangement in the naked mole rat lenses, especially at the posterior margin in 15- to 30-day-old juveniles, is abnormal. The newly differentiated lens fibers, which do not express γ-crystallins, do not form concentric circles around the embryonic nucleus, as in the mouse, but are predominantly located in the posterior hemisphere of the lens (Fig. 9B–D). We think that such an irregular arrangement of lens fibers could result from displacement of the equatorial region. A consequence of this displacement is that the elongating secondary lens fibers are forced to grow into the posterior lens hemisphere, rather than to elongate equally in both the anterior and the posterior directions. The embryonic lens nucleus is progressively pushed toward the lens epithelium, rather then remaining in the center of the lens, with the result that the lens fibers become organized into the irregular bent arrangement observed in the naked mole rat (Fig. 10).
In summary, we conclude that the naked mole rats provide a potentially very interesting model for investigations into the developmental and genetic basis of evolutionary change and eye disease. However, the usefulness of this animal as the model for human eye disease is limited by their slow reproduction rate and the fact that the embryos are not easily available.
We thank Melaney Peterson for assistance with histological sectioning and TEM, Toni Wiggins for technical support, Liz van der Merwe for the SEM micrograph of the adult mouse ciliary body and help with TEM sectioning, and Barbara Young for TEM sectioning. We also thank Dr. Linlin Ding (Horwitz Laboratory, Jules Stein Eye Institute, University of California, Los Angeles, CA) for providing us with anti-γ-crystallin antibodies, Thandi Mgwebi for allowing us to use her SEM micrograph of the adult mouse corneal endothelium, as well as Dr. Dirk Lang for helpful discussions. This work was supported by grants from the Medical Research Council of South Africa (to S.H.K.) and National Research Foundation (to M.J.O.) and by bursaries from the Medical Research Council of South Africa and the University of Cape Town (to N.V.N.).