| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
From the Department of Ophthalmology, Columbia University, New York, New York.
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResultsDiscussionReferences |
|---|
METHODS. A transgenic mouse line that was engineered to ubiquitously express green fluorescent protein (GFP) was used to visualize corneal epithelial cells. For histology, epithelial GFP was imaged in a wholemounted cornea en face, and also in frozen cross-sections, under a fluorescence microscope. For in vivo studies, the anesthetized mouse was placed on a custom-made observation platform under a fluorescence microscope. Epithelial fluorescence was digitally recorded two to three times a week, and a rate of cell movement was determined from the time-lapse sequences.
RESULTS. The GFP expression in the corneal epithelium was nearly ubiquitous up to about 1 week after birth, and thereafter it gradually became sporadic, resulting in a mosaic pattern of GFP positive cells, with the brightest cells present in the basal and suprabasal layer of the epithelium. Both high- and low GFP-cells formed radial streaks toward the central cornea, frequently displaying vortex patterns at the center. Clusters of several high-GFP cells were tracked in living mice for up to 7 weeks, and an analysis of time-lapse sequences revealed that they moved centripetally at an average rate of 26 µm/d.
CONCLUSIONS. Corneal epithelium of adult GFP mice exhibits a pattern of GFP expression that is suitable for studying cell movement in the normal cornea. Epithelial cells at the basal or suprabasal layers move centripetally in these mice at an average rate of 26 µm/d.
Cell migration is one of the most fundamental aspects of epithelial homeostasis, and there has been a steady increase in our knowledge in this area. A strong body of evidence now suggests that corneal epithelial cells arise from the stem cells at the limbus.1 2 3 4 5 6 Once inside the cornea, epithelial cells are thought to move slowly toward the center, as delineated by the X, Y, Z hypothesis of Thoft and Friend.7 Evidence in support of this concept has been provided by various clinical observations.8 9 10 11 In addition, a direct observation in humans was reported by Auran et al.,12 who showed, using specular confocal microscopy, that some basal epithelial cells migrate centripetally, 23, 29, and 32 µm over a 24-hour period in three measurements in one eye. However, it remains to be determined whether these short, limited observations can be expanded to show long-term movement of epithelial cells.
Studies in the animal have also been instructive. Kinoshita et al.13 presented the first experimental evidence of centripetal movement of cells after lamellar keratoplasty in the rabbit, although their experiments did not address cell movement in the normal cornea directly. Buck14 was the first to demonstrate the centripetal movement of epithelial cells in the normal cornea. He labeled epithelial cells of the mouse cornea with India ink and determined the rate of movement to be approximately 17 µm/d. However, the observations lasted only 7 days, and the total measured distance was approximately 120 µm, which is less than 10% of the radius of a typical mouse cornea. Thus, it is unclear whether this short-term, short-distance observation can be extrapolated to the general movement of epithelial cells. Also uncertain is whether the labeling with India ink altered the behavior of the cells.
Thus, despite the evidence suggesting centripetal movement of epithelial cells in the normal cornea, there have been only two studies that presented a direct observation of such movement.12 14 As such, many basic questions remain unanswered, such as whether the movement is continuous or intermittent, whether it is operative in all areas of the cornea, which layer of epithelium moves at what rate, what the driving force is, and how migration and mitosis are coordinated.
There has not been much investigation concerning movement of epithelial cells in the normal cornea, presumably because no suitable methodology was available to examine slow movement of cells. Histologic studies are not always appropriate for investigating dynamic events such as cell movement. Accordingly, we sought to establish an animal model in which the migration of epithelial cells could be studied in the uninjured, normal cornea. We report that, by using the GFPU mouse in conjunction with the in vivo microscopy system we developed previously, we were able to observe directly the centripetal movement of corneal epithelial cells for up to 7 weeks in a living mouse. This experimental system should be valuable for further studies on epithelial cell migration in the normal cornea.
|
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Histology
The eyes of GFPU mice at various ages were isolated after the animal was killed with pentobarbital (100 mg/kg), fixed in formaldehyde, and stained with the nuclear fluorescence dye 4,6-diamidino-2-phenylindole (DAPI), and the corneal size determined. For this, the eyeball was placed under an upright fluorescence microscope (Axioskop2; Carl Zeiss, Oberkochen, Germany) with the cornea facing up. The DAPI image was acquired digitally (Orca, Hamamatsu, Japan) and analyzed on computer (Photoshop; Adobe Systems, San Jose, CA). The corneal boundary was defined as the line across which nuclear density increases sharply, indicating the epithelial cell transition from cornea to limbus. The corneal size was determined by measuring the distance of two perpendicular chords of the cornea and averaging them. It should be noted that this is not surface distance on the cornea but rather the longest linear distance between two margins of the cornea that is naturally situated in the eye.
Epithelial GFP was digitally recorded with the fluorescence microscope after dissecting the cornea from an eye and making several radial cuts so that it could be mounted flat.
Some corneas were snap frozen, and 8-µm thick cross-sections were prepared with a cryostat (CM1850; Leica, Wetzlar, Germany). The cryosections were stained with DAPI and digitally recorded with the fluorescence microscope for both DAPI and GFP.
In Vivo Microscopy
Technical details of in vivo microscopy of living mouse cornea will be reported elsewhere (Maurice D, Zhao J, Nagasaki T, manuscript in preparation). The microscope for the histology, a standard upright fluorescence microscope, as described earlier, was used with a 40x water-immersion (numerical aperture [NA], 0.8), a 10x dry (NA, 0.25), and a 4x dry (NA, 0.1) objective. Mice were anesthetized with isoflurane in oxygen. Anesthesia was maintained by supplying steady flow of anesthetic gas to the nose of the mouse through a nose cone. Before microscopic observation, topical xylazine (20 mg/mL in saline) was given to stimulate proptosis of the eye,16 so that a wide area of the corneal surface, including the limbus, could be observed without forcefully opening the eyelid. In some mice, when the pupil became dilated, 1% pilocarpine was given topically to promote miosis. This was necessary because strong fluorescence from lens GFP interfered with the observation of corneal fluorescence.
To reduce motion blur introduced by the mouse’s breathing, its head was held and immobilized with an upper-jaw clamp, which was then magnetically secured to a mouse holder so that the cornea was positioned directly under, and facing toward, the objective lens. When the 40x objective was used, the cornea was lightly touched with a glass coverslip at the bottom of a conical spacer, which was secured on the microscope’s stage, to suppress eye movements in the z direction. The glass applanation also helped flatten the cornea so that a nearly entire field of observation (430 x 340 µm after digitization) was in focus. The mouse was placed on a platform attached to custom-made gimbals that could be rotated freely in all directions, which facilitated rapid scanning of a wide area of the cornea (Maurice D, Zhao J, Nagasaki T, manuscript in preparation).
Microscopic images were captured with a digital camera (Orca) through a relay zoom lens (0.4x to 2.0x; Carl Zeiss). To minimize phototoxicity to live cells, an illumination source (100-W Hg arc lamp) was operated at 50% or less of its full power. A narrow band-pass excitation filter (480 ± 20 nm; Chroma Technology, Brattleboro, VT) also helped to reduce unnecessary illumination. In addition, a mechanical shutter for the arc light (Uniblitz; Vincent Associates, Rochester, NY) was used to ensure that corneal exposure to light was minimal. The camera and the shutter were synchronously controlled by an image-processing package (Metamorph; Universal Imaging Corp., West Chester, PA), which was also used for postacquisition image processing.
For routine image acquisition, a few images from different angles were taken with a 4x objective with 0.5x zoom , and 5 to 15 overlapping microscopic fields with a 10x objective with 0.5x zoom. In some animals, 10 to 20 overlapping fields were recorded with a 40x objective with 0.5x zoom to document the details of the corneal surface, which allowed analysis of individual epithelial cells. However, the 40x objective was used only sparingly to minimize the possibility of illumination damage to epithelial cells. A wide-field view was reconstructed from the overlapping images (Photoshop; Adobe). With 10x images, only the central one half to one third of each image was used, because the peripheral zone was out of focus. This assured that a reconstructed image was suitable for measurement of actual corneal surface distance. Under these conditions, image resolution was approximately 3.4, 1.3, and 0.33 µm/pixel for images acquired with a 4x, 10x, and 40x objective, respectively.
Cell-Tracking and Analysis
The cell movement was analyzed on computer (Photoshop; Adobe) from reconstructed images taken with either a 10x or a 40x objective. A GFP-positive epithelial cell cluster with a unique geometric pattern was chosen for tracking. A composite image of the recorded area was displayed on a computer monitor and either a leading edge or a trailing edge of the chosen cluster was marked and saved as an overlay image. Most of the limbal capillary vessels were strongly GFP positive and served as a fixed reference marker. Two or more marker locations were defined from uniquely shaped capillary branches at an area of limbus nearest to the epithelial cells to be tracked and saved in the same overlay image. The epithelial movement was measured in reference to one of these marker points. When the main marker was not identified, its location was estimated from the other markers.
|
Results |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
To investigate whether the level of expression was dependent on development, we looked at the epithelial level of GFP as well as the size of the cornea in the GFPU mice at various ages. The nominal size of the cornea doubled in the first 2 weeks after birth, expanding more than 100 µm/d, and reaching the maximum size at 8 to 10 weeks (Fig. 1) .
|
|
|
Examination of fixed corneas with high-GFP cells revealed that some of these cells in the basal epithelium were going through mitosis (Fig. 4) , indicating that the high-GFP cells were competent for mitosis and that their progeny were likely to maintain a high-GFP phenotype.
|
|
|
|
|
|
To evaluate the cell movement quantitatively, the location of selected high-GFP cell clusters was determined at various time points in reference to a fixed marker at the limbus in seven recordings from four corneas (Fig. 10) . The plots show that the rate of cell movement was fairly constant throughout the cornea, from the periphery to the central zone. Rates of movement were determined from these plots (Table 1) , and the average rate of centripetal movement was found to be approximately 26 µm/d. The steady rates of movement over several weeks suggest that any illumination-induced phototoxicity was minimal.
|
|
|
Discussion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Expression of GFP in the young mouse, 4 weeks or younger, was not stable, but in older animals, expression levels of GFP in limbal stem cells seemed to be generally maintained in daughter cells, as evidenced by the presence of radial stripes that emerged from the limbus (Fig. 2) . Furthermore, our results in Figure 7 indicate that a cluster of high-GFP cells remained high in GFP over 6 weeks, which would be long enough for a few generations of these cells. These observations, though indirect, suggest that the limbal stem cells express different levels of GFP and that this trait is inherited by the daughter cells and further by their progeny, at least in the 7- to 15-week-old adult mice that we used for measurement of centripetal movement of cells.
Centripetal Movement of High-GFP Cells
Radial patterns of epithelial GFP in histologic specimens (Fig. 2) first gave a hint of centripetal movement of cells, similar to many clinical observations.9 11 Similar patterns were observed by Collinson et al.17 in lac-Z chimeric mice, which exhibited dramatic and unmistakable stripes. The radial patterns were not obvious, however, during the first 2 weeks after birth, when the cornea grew at a rate of more than 100 µm/d under the closed eyelid (Fig. 1) . It seems that centripetal movement of cells in a direction opposite that of corneal growth was completely negated by the rapid growth of the cornea. When the radial patterns became well established in the adult mouse, aged approximately 8 weeks and more, high-GFP cell clusters were generally more concentrated in the central portion of the cornea, which is consistent with their centripetal movement and convergence in the center.
In vivo microscopy provided a more direct and convincing evidence of centripetal movement of cells (Figs. 7 8 10) . The rate of movement was approximately 26 µm/d (Table 1) , which is comparable to the previous findings in the mouse14 and the human12 (see the introduction).
Vortex Pattern of GFP
Some of the corneas exhibited vortex patterns in the central cornea, which were also present in the lac-Z chimeric mice,17 but how they came to exist is not clear. Developmentally, appearance of central vortex patterns coincided with that of radial stripes at the peripheral cornea, approximately 4 weeks after birth, suggesting that centripetal movement of cells may be involved in the formation of the vortex. Thus, it is plausible that the vortex is a natural consequence of many cells converging in a small central area from all directions, as originally suggested by Bron.9 If this is true, a constant migration cue, or a driving force, in the centripetal direction should be assumed, possibly a population pressure generated at a peripheral zone of the cornea. However, there were some corneas in which no vortex could be identified, despite the unequivocal presence of radial stripes (data not shown), suggesting that formation of the vortex may require more than a simple pressure from the periphery.
In humans, the whorl pattern was almost always found to be clockwise, and electromagnetic fields of the eye have been suggested as the cause.18 However, the whorls in the GFPU mouse cornea were both clockwise and counterclockwise, similar to those in the Lac-Z chimeric mice,17 and therefore the electromagnetic theory does not seem applicable to these mice.
Mitosis and the Centripetal Movement of Cells
We observed movement of high-GFP cells from limbus to central cornea over 7 weeks (Figs. 7 10) . If the turnover time of epithelial cells was 2 weeks,19 there would have been at least three generations of high-GFP cells during this period. A shorter generation time20 21 would have required more frequent mitoses within the high-GFP cell cluster. Therefore, the observed high-GFP cells during the 7-week period were not the same cells, but they collectively represented the progeny of high-GFP basal cells that were present at the beginning of the recording. The occurrence of multiple mitoses within the high-GFP cell cluster indicates that the observed centripetal movement, averaging approximately 26 µm/d (Table 1) , was a combined result of cell division and migration.
The size of high-GFP cell clusters, thus the number of these cells, appeared unchanged throughout their movement for a long distance. This suggests that every mitosis resulted in, on the average, one surviving high-GFP cell, which is in agreement with an earlier suggestion.21 In this context, it is interesting to note that postmitotic cells were almost always found as a pair of daughter cells in proximity to each other in both basal and wing cell layers.22 If this were true of high-GFP cells, it would suggest that one pair of high-GFP daughter cells survives after every two mitoses, instead of one daughter cell surviving from a single mitosis. The other pair of high-GFP daughter cells were probably lost from the surface rapidly, so as to maintain the size of the high-GFP cell cluster.
Alternatively, a possibility exists that all the high-GFP cells that we tracked (Table 1) were replication incompetent and that the observed high-GFP cells were in fact the same epithelial cells throughout the recordings. We think this to be unlikely, however, because our results demonstrated that high-GFP cells could go through mitosis (Fig. 4) .
Driving Force
The driving force of the centripetal movement of cells is not known, although several proposals have been advanced. There are four viable published hypotheses, which are not mutually exclusive: (1) preferential desquamation of central corneal epithelium, drawing peripheral cells toward the central cornea11 23 ; (2) population pressure from limbus and peripheral cornea due to proliferation and immigration of cells4 9 24 ; (3) a gradient of chemical signals emanating from limbal capillary vessels14 ; and (4) stimulation by epithelial sympathetic nerves.25
Although our observations did not provide any direct evidence, they suggest that population pressure plays an important role. We observed fine vortex patterns at the center of the normal cornea (Fig. 3) , which we believe is best explained by the population pressure theory discussed earlier. We also observed that the general shape of high-GFP cell clusters and their positions relative to each other were maintained while they moved centripetally (Figs. 7 8) . This may be an indication that there was a tightly regulated external force, such as that provided by population pressure, that positively guided the cells for directional movement. A source of population pressure can be cell division and/or cell migration. However, some have claimed a higher mitotic rate in the peripheral cornea compared with the central cornea,17 20 26 27 28 whereas others have reported similar rates between them,23 29 30 31 32 and consequently the contribution of cell division remains unclear. If the latter possibility is true, population pressure must be provided by continual migration of cells from the limbus into the cornea.
Regardless the role of population pressure, the other three hypotheses remain a possibility at this time. In particular, we are attracted by a stromal molecular gradient originating from the limbal capillaries33 as a possible contributor to the centripetal movement of cells.14
Homeostasis of Normal Corneal Epithelium
Although dynamics of cell renewal in the stratified epithelium is complex, a relatively simple picture emerges from the results of earlier investigations as well as the present study. As an extension of the X, Y, Z theory,7 we envisage five major components in the epithelial homeostasis of normal cornea: (1) generation of transient amplifying cells from the stem cells and their entrance into the cornea1 2 3 4 5 6 ; (2) cell divisions at the basal epithelium19 26 29 ; (3) vertical movement of basal cells and their daughter cells toward the surface, which is associated with terminal differentiation4 20 21 22 34 ; (4) horizontal movement of cells toward the central cornea7 12 13 14 (present study); and (5) loss of surface cells by exfoliation.35
Basal cell division and vertical movement of cells have been well characterized, but the other three are relatively poorly understood. In this study, the centripetal movement of epithelial cells occurred at a steady rate of approximately 26 µm/d (Fig. 10 , Table 1 ), for almost an entire distance of the corneal radius in one case (Fig. 7) . The steady centripetal movement would cause overcrowding in the central cornea unless cells were removed as rapidly as they accumulate.7 In fact, experimental evidence suggests that surface cells are lost preferentially in the central cornea.35 36 Although the role of central desquamation in centripetal movement of cells is not certain, it is clearly an essential process for epithelial homeostasis.
The steady rate of centripetal movement of cells also suggests that the limbal stem cell progeny enters the cornea at a similar rate, although this has not been observed directly in the normal eye. Consequently, little is known about the nature of this immigration process, such as rate, frequency, and location. We believe it is now possible to investigate these parameters with the technique used in this study by continuously tracking the movement of high-GFP cells in the limbus.
The question now is how these distinct events are coordinated or how each of the five components influences the outcome of others, for the common goal of maintaining the epithelial mass. A mathematical model24 may be a useful approach if the parameters for these components could be determined experimentally. Some pathologic conditions, such as persistent defects and wound healing may be better understood by examining these five components separately.
In summary, the GFPU mouse offers a unique opportunity to study movement of epithelial cells in the normal cornea. Further microscopic analyses of these corneas, combined with various biochemical markers of epithelial differentiation,4 37 should be useful in understanding epithelial homeostasis.
|
Acknowledgements |
|---|
|
Footnotes |
|---|
Submitted for publication July 12, 2002; revised September 9, 2002; accepted September 17, 2002.
Disclosure: T. Nagasaki, None; J. Zhao, None
The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked "advertisement" in accordance with 18 U.S.C. 
1734 solely to indicate this fact.
Corresponding author: Takayuki Nagasaki, Department of Ophthalmology, Columbia University, 630 W 168th St., New York, NY 10032; tn4{at}columbia.edu.
|
References |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  C. Q. Yu, M. Zhang, K. I. Matis, C. Kim, and M. I. Rosenblatt Vascular Endothelial Growth Factor Mediates Corneal Nerve Repair Invest. Ophthalmol. Vis. Sci., September 1, 2008; 49(9): 3870 - 3878. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. V. Patel and C. N. J. McGhee In Vivo Laser Scanning Confocal Microscopy Confirms that the Human Corneal Sub-basal Nerve Plexus Is a Highly Dynamic Structure Invest. Ophthalmol. Vis. Sci., August 1, 2008; 49(8): 3409 - 3412. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. K. Swamynathan, J. Davis, and J. Piatigorsky Identification of Candidate Klf4 Target Genes Reveals the Molecular Basis of the Diverse Regulatory Roles of Klf4 in the Mouse Cornea Invest. Ophthalmol. Vis. Sci., August 1, 2008; 49(8): 3360 - 3370. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Dvorscak and C. F. Marfurt Age-Related Changes in Rat Corneal Epithelial Nerve Density Invest. Ophthalmol. Vis. Sci., March 1, 2008; 49(3): 910 - 916. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Z. Li, A. R. Burns, R. E. Rumbaut, and C. W. Smith {gamma}{delta} T Cells Are Necessary for Platelet and Neutrophil Accumulation in Limbal Vessels and Efficient Epithelial Repair after Corneal Abrasion Am. J. Pathol., September 1, 2007; 171(3): 838 - 845. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. K. Swamynathan, J. P. Katz, K. H. Kaestner, R. Ashery-Padan, M. A. Crawford, and J. Piatigorsky Conditional Deletion of the Mouse Klf4 Gene Results in Corneal Epithelial Fragility, Stromal Edema, and Loss of Conjunctival Goblet Cells Mol. Cell. Biol., January 1, 2007; 27(1): 182 - 194. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. H. Levin and A. S. Verkman Aquaporin-3-Dependent Cell Migration and Proliferation during Corneal Re-epithelialization. Invest. Ophthalmol. Vis. Sci., October 1, 2006; 47(10): 4365 - 4372. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. Adachi, H. Ulanovsky, Y. Li, B. Norman, J. Davis, and J. Piatigorsky Serial Analysis of Gene Expression (SAGE) in the Rat Limbal and Central Corneal Epithelium. Invest. Ophthalmol. Vis. Sci., September 1, 2006; 47(9): 3801 - 3810. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Zhou, X.-m. Li, and R. M. Lavker Transcriptional Profiling of Enriched Populations of Stem Cells Versus Transient Amplifying Cells: A COMPARISON OF LIMBAL AND CORNEAL EPITHELIAL BASAL CELLS J. Biol. Chem., July 14, 2006; 281(28): 19600 - 19609. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. V. Patel and C. N. J. McGhee Mapping of the Normal Human Corneal Sub-Basal Nerve Plexus by In Vivo Laser Scanning Confocal Microscopy Invest. Ophthalmol. Vis. Sci., December 1, 2005; 46(12): 4485 - 4488. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. G. Every, J. P. Leader, A. C. B. Molteno, T. H. Bevin, and G. Sanderson Ultraviolet Photography of the In Vivo Human Cornea Unmasks the Hudson-Stahli Line and Physiologic Vortex Patterns Invest. Ophthalmol. Vis. Sci., October 1, 2005; 46(10): 3616 - 3622. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Nagasaki and J. Zhao Uniform Distribution of Epithelial Stem Cells in the Bulbar Conjunctiva Invest. Ophthalmol. Vis. Sci., January 1, 2005; 46(1): 126 - 132. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. M. Collinson, S. A. Chanas, R. E. Hill, and J. D. West Corneal Development, Limbal Stem Cell Function, and Corneal Epithelial Cell Migration in the Pax6+/- Mouse Invest. Ophthalmol. Vis. Sci., April 1, 2004; 45(4): 1101 - 1108. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. I. Shestopalov and S. Bassnett Development of a macromolecular diffusion pathway in the lens J. Cell Sci., October 15, 2003; 116(20): 4191 - 4199. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||
