Uniform Distribution of Epithelial Stem Cells in the Bulbar Conjunctiva

Abstract

purpose. To determine the distribution of epithelial stem cells in the bulbar conjunctiva by measuring homeostatic movements and mitosis of epithelial cells in this region.

methods. The ubiquitous GFP mouse was used to monitor movement of conjunctival epithelial cells. Cell movement was determined by histology, analyzing the shape and distribution of GFP cell clusters in flat wholemount specimens, and by in vivo time-lapse microscopy, tracking the movement of GFP-positive cells in the bulbar conjunctiva near the limbus. Mitoses were determined by labeling DNA of adult mice with bromodeoxyuridine (BrdU) for 3 days. Label-retaining cells (LRCs) were determined by a pulse label of newborn mice with BrdU, followed by a chase of 6 to 7 weeks.

results. Similar to the corneal epithelium, only some of the conjunctival epithelial cells expressed a high level of GFP. Histology showed that GFP-positive cells existed as clusters of several to several dozen cells. No stripe pattern of GFP was observed in any part of the conjunctiva, suggesting that directed cell movement was rare or nonexistent. Time-lapse analyses revealed that none of the tracked GFP clusters exhibited a continuous and directed movement and that most GFP clusters were stationary for several weeks and much longer in some occasions. BrdU labeling showed that GFP-positive cells in this region were mitotically active. BrdU pulse–chase experiments demonstrated that LRCs were distributed uniformly in this region.

conclusions. Epithelial cells of the bulbar conjunctiva near the limbus are mitotically active and yet they are generally immobile in a lateral direction, indicating that these cells are self-sufficient. These results, combined with the uniform distribution of LRCs, suggest that epithelial stem cells are distributed uniformly in this area.

Conjunctival epithelium is a self-renewing tissue with rapid cell turnover, and its stem cells are thought to be present within the tissue, supplying differentiated epithelial cells throughout the lifetime of a host body. ¹ Location of conjunctival epithelial stem cells has been controversial, and various parts of the tissue have been suggested as an area of concentrated stem cells, including limbus (rat ² ), bulbar conjunctiva (human ³ ), fornix (rabbit, ⁴ mouse, ^{5 6} and human ³ ), palpebral conjunctiva (rat ⁷ ), and mucocutaneous junction (rat ² and rabbit ⁸ ). A clinical observation indicated that conjunctival stem cells are in the fornix and/or bulbar conjunctiva. ⁹  

Techniques to estimate the location of stem cells included (1) determination of slow-cycling cells (label-retaining cells [LRCs]) by mitotic DNA labeling (fornix, ^{5 8} palpebral conjunctiva, ⁷ and mucocutaneous junction ⁸ ), (2) mitotic stimulation by phorbol ester (fornix ⁶ and palpebral conjunctiva ⁷ ), (3) proliferative capacity of isolated cells in culture (fornix ^{3 4} and bulbar conjunctiva ³ ), and (4) determination of an origin of cell movement (limbus ² and mucocutaneous junction ⁸ ). These results are not necessarily in conflict, because conjunctival stem cells may indeed be distributed throughout the conjunctiva. It seemed clear, however, that these results, obtained with a variety of techniques in different species, should be evaluated carefully with additional experiments to determine the precise location and distribution of epithelial stem cells within the conjunctiva. 

Examination of homeostatic cell movement is a potentially powerful method for identifying a stem cell location directly, because the origin of cell movement would be the exact area of stationary stem cells and their niches. Three published reports ^{2 8 10} in which this technique was used relied on an indirect approach by estimating cell movement from metabolic DNA labeling in tissue cross sections, where interpretation of results is not necessarily straightforward. Thus, Zajicek et al. ¹⁰ reported bulbar conjunctival epithelial movement to be 10.5 ± 2.4 μm/d from the limbus to the fornix in rats. Subsequently, the same group ² reported that both bulbar and palpebral conjunctival cells move toward the fornix at 13.2 and 11.8 μm/d, respectively. Wirtschafter et al. ⁸ concluded that rabbit conjunctival epithelial cells of the mucocutaneous junction move toward the fornix at 1.7 mm/d, which seems extraordinarily fast. The only direct measurement of conjunctival epithelial cells was reported by Buck, ¹¹ who stated, without providing data, that conjunctival epithelium close to the limbus had moved neither toward nor away from the limbus 7 days after they were labeled with India ink. As such, the existing data on conjunctival epithelial cell movement are in dispute. Therefore, we initiated this study to measure directly the movement of bulbar conjunctival epithelial cells with an in vivo time-lapse microscopy technique, ^{12 13} in an effort to determine stem cell location in this area. 

Materials and Methods

Animals

Animal studies adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. GFPU mice ¹⁴ were obtained from the Jackson Laboratory (Bar Harbor, ME) and maintained in house. Only male mice were used in this study. 

Histology

For histology, eyes of GFPU mice at various ages were isolated after the animal was killed with an intraperitoneal injection of pentobarbital (100 mg/kg). To prepare flat wholemounts containing an entire area of the ocular surface, we removed the intact globe together with eyelids by first cutting the skin just outside of both lid margins and then dissecting connective tissues around the globe (Figs. 1A 1B) . After enucleation, eyelids were turned inside out (Figs. 1C 1D 1E)to expose the entire ocular surface, from the cornea to the mucocutaneous junction (Fig. 1E) . The exposed eye was fixed in formaldehyde and dissected, to prepare a flat wholemount of the ocular surface, by first cutting the anterior segment into two pieces (superior and inferior) and then making several radial cuts in the cornea and conjunctiva so they could be mounted flat. Muscle and other accessory tissues were carefully removed with scissors. Wholemounts were stained with a nuclear fluorescence dye, 4,6-diamidino-2-phenylindole (DAPI), and both DAPI and green fluorescent protein (GFP) patterns were imaged under a fluorescence microscope (Axioskop2; Carl Zeiss Meditec, Oberkochen, Germany). Overlapping microscopic images were acquired digitally (Orca 100; Hamamatsu, Hamamatsu City, Japan, and Metamorph; Universal Imaging Corp., Downingtown, PA), and assembled on computer (Photoshop; Adobe Systems, Mountain View, CA) to prepare a large image file of the entire ocular surface. 

Goblet cells were identified in flat wholemounts by the presence of large vacuoles that pushed the nucleus to the bottom of a cell and the cytoplasm to the periphery, as visualized by phalloidin staining ¹⁵ (phalloidin-Alexa 546; Molecular Probes, Eugene, OR) and by cytoplasmic distribution of GFP in GFP-positive cells. Some eyes were snap frozen in liquid nitrogen after exposure of the ocular surface, and 8-μm-thick cross sections were prepared with a cryostat (CM1850; Leica, Heidelberg, Germany). The cryosections were stained with DAPI and digitally recorded with the fluorescence microscope for both DAPI and GFP. 

In Vivo Microscopy

In vivo microscopy and digital imaging were performed as described previously. ^{12 13} Mice were anesthetized with 3% isoflurane in oxygen, supplied as a steady flow of gas to the nose of the mouse. The right eye was lightly proptosed with a thin vinyl-coated U-shaped metal wire to expose an inferior temporal quadrant of the bulbar conjunctiva for microscopy. The wire made contact with parts of conjunctiva away from the area of observation, but this apparently did not influence cell movement in the area of observation. The area of recording was centered at the inferior temporal bulbar conjunctiva, because other areas could not be fully exposed and brought under the objective for in vivo microscopy due to the arrangement of the mouse holder. 

For acquisition of GFP fluorescence images, several overlapping microscopic fields were captured with a 5× objective (Fluar; Carl Zeiss Meditec) with a telephoto zoom lens set at 0.5×. Image capture was complete within 1 minute, and the entire process took <5 minutes from the start of anesthesia induction. A wide-field view was constructed from the overlapping images (Photoshop; Adobe Systems). Image resolution was approximately 2.6 μm/pixel under these conditions. GFP patterns were recorded at least once a week, and the movement of GFP-positive cells was analyzed from time-lapse sequences. 

The size of GFP-positive cell clusters was determined from time-lapse images by first manually drawing cluster margins with Photoshop (Adobe Systems), and then measuring the area with Metamorph (Universal Imaging Corp.). To determine the movement of GFP clusters, 90 clusters from 10 time-lapse recordings were selected because they could be continuously identified for at least four consecutive weeks. X-y coordinates of each cluster’s centroid were determined, using unique limbal capillary branches as fixed reference markers. Cluster movement was tracked at each time point to determine the number of clusters that moved in one direction for at least 200 μm over a 3-week period (>9.5 μm/d). Also determined was the number of clusters whose net movement was <200 μm over an 8-week period (<3.6 μm/d). The shape of each cluster was monitored to detect the movement of constituent cells, but this information was not used to determine the mobility of a cluster itself as a whole. 

Labeling of DNA with BrdU

DNA was metabolically labeled with bromodeoxyuridine (BrdU; Sigma-Aldrich, St. Louis, MO) in live mice to determine the presence of cells that are synthesizing DNA and LRCs. ^{16 17} For the DNA synthesis detection, BrdU was injected intraperitoneally in adult mice at 50 mg/kg every 12 hours for 3 days (total of six injections). The mice were killed 12 hours after the last injection and wholemounts prepared as described earlier. BrdU was detected immunohistochemically with rat anti-BrdU antibody (Serotec Inc., Raleigh, NC) and donkey anti-rat IgG-Cy3 conjugate (Jackson ImmunoResearch Laboratories, West Grove, PA). For LRC detection, BrdU was injected into 1-day-old mice subcutaneously at 20 mg/kg every 12 hours for 4 days (total of eight injections). The mice were kept as usual for 6 to 7.3 weeks (a chase period), and killed for BrdU histology, as just described. BrdU-positive cells (LRCs) were identified under the fluorescence microscope and their positions manually plotted in an entire area of the bulbar conjunctiva up to approximately 1 mm from the limbus. An area beyond that point was not scored due to occasional tissue folding which made it excessively laborious to distinguish epithelial cells from stromal cells and goblet cells. 

Results

Cell Movement by Histology

To investigate the conjunctival epithelial cell movement, we used GFPU mice ¹⁴ which had been shown to be valuable in studying corneal epithelial cell movement. ¹² We developed a technique to prepare a flat wholemount tissue that contained both cornea and conjunctiva in their entirety (Fig. 1) , so that the entire ocular surface could be studied simultaneously. Examination of the wholemounts revealed that, similar to the corneal epithelium, ¹² not all conjunctival epithelial cells were strongly GFP positive (Fig. 2) . As reported previously, ¹² young corneas contained mosaic GFP patterns, and adult corneas exhibited radial GFP stripes from the limbus to the central cornea. In contrast, conjunctival GFP-positive cells formed small patches, and there was no striped pattern in any of the >20 wholemount specimens of various ages we examined (4-, 14-, and 32-week-old specimens are shown in Fig. 2 ). GFP-positive epithelial patches were found as cell clusters of various shapes containing several to several dozen cells. The cluster size ranged from approximately 50 to several hundred micrometers (Fig. 3) . 

The fornix contained the brightest GFP patches (Figs. 2 3)due to an abundance of goblet cells that generally contained a higher level of GFP than epithelial cells did (data not shown). Occasionally, wholemount preparations contained muscle cells in the fornix and palpebral regions, because it was difficult to separate them from the conjunctiva without destroying its sheet structure. Muscle cells were always much brighter with GFP than epithelial cells or goblet cells and were clearly identifiable in the wholemounts under the microscope because of their focal plane’s being farther beneath the conjunctival stroma. 

Distribution of GFP-positive epithelial cells appeared to be random. We did not find any preferential localization of GFP-positive cells in any part of the conjunctiva. After examining the entire limbal circumference from more than 20 adult eyes, we failed to detect any correlation between the corneal GFP stripes and the conjunctival GFP clusters, suggesting that corneal GFP cells and conjunctival GFP cells arose independently. 

Cryosections demonstrated that the bulbar conjunctival epithelium near the limbus (the area of in vivo microscopy; described in the next section) was made up of roughly two cell layers (Fig. 4B) . GFP-positive cells were found in both basal and superficial layers of the epithelium, but GFP fluorescence in the stroma and other structures was minimal (Fig. 4A) . 

Phalloidin staining ¹⁵ identified no goblet cells (Fig. 5F)in the bulbar conjunctiva up to approximately 800 μm from the limbus. An increasing number of goblet cells were present beyond the zone of 800 μm toward the fornix (Fig. 5C) . When goblet cells were GFP positive, cytoplasmic GFP distribution clearly identified them by the large vacuole that excluded GFP (Fig. 5B) . Using this criterion, we reaffirmed that there were no goblet cells inside the GFP-positive cell clusters within 800 μm of the limbus. 

Cell Movement by In Vivo Microscopy

Because the inferior temporal bulbar conjunctiva near the limbus was accessible and therefore suitable to in vivo microscopy, we investigated the cell movement in this area in a live mouse. GFP-positive cells in this region formed clusters with a unique geometric shape, which assisted relocation, allowing us to track them over time. Only those clusters that were within 600 μm of the limbus were tracked, to ensure that goblet cells were not included. The area of 143 randomly selected clusters (not all of them were tracked) from 10 time-lapse recordings ranged from 0.001 to 0.114 mm², with the average being 0.014 ± 0.018 mm² (mean ± SD), which is likely to be an overestimation, because a large cluster may consist of multiple small clusters. 

In the representative time-lapse sequence shown in Figure 6 , several GFP clusters were observed to stay immobile between 17 and 21.3 weeks. At some time between 21.3 and 21.6 weeks, many of them changed positions and/or shape drastically, so that tracking of individual clusters became impossible. The movement did not continue, however, because the positions of the clusters at 21.6 weeks were maintained for approximately 30 weeks thereafter with a minimal change in shape, indicating that there was no directed cell movement in this area. Time-lapse images suggested that the largest movement over the 30 weeks was approximately 300 μm, which translates to a rate of <2 μm/d. As a comparison, centripetal movement of corneal epithelial cells was determined to be approximately 26 μm/d. ¹²  

When we analyzed movement of 90 randomly selected GFP clusters in 10 time-lapse sequences, we found that none of the tracked clusters exhibited a unidirectional movement of at least 200 μm in 3 weeks (Table 1) . Movement, if any, was generally less than a few micrometers per day without a fixed direction. An exception to this was a sudden, one-time movement that occurred occasionally, such as that seen in Figure 6 , between 21.3 and 21.6 weeks. The cause of such movement may be an external physical insult to the area of observation, such as accidental scratching, rather than a physiological one. A further analysis showed that approximately 70% of the tracked clusters were nearly immobile for at least eight consecutive weeks (Table 1) . The age of the animals did not seem to play a major role between 11 and 53 weeks. Although patches of GFP cell clusters themselves were immobile, a small and continuous shape change of individual clusters was observed consistently, most likely due to local movements and/or turnover of individual cells within the cluster, which we did not analyze in this study. 

Mitotic Rates

To determine whether GFP-positive cells were mitotically active or quiescent, we examined DNA synthesis of epithelial cells in the area of in vivo microscopy with metabolic labeling of DNA with BrdU. The results showed that DNA synthesizing cells were distributed uniformly in this region, and no correlation was apparent between DNA synthesis and GFP expression (Fig. 7) . This strongly suggests that the GFP-positive stationary cells that were recorded with in vivo microscopy (e.g., Fig. 6 ) contained mitotically active cells. 

LRC Distribution

Most of the LRCs ^{16 17} are likely to be stem cells in the conjunctival epithelium, where cells turn over continuously. ⁵ Therefore, to estimate the stem cell distribution, we determined the presence of LRCs in the conjunctival region used for in vivo time-lapse microscopy. As shown in Figure 8 , LRCs were distributed uniformly in the bulbar conjunctiva, suggesting that stem cells are distributed uniformly in this area. Quantitation of LRCs in the bulbar conjunctiva within 0.8 mm of the limbus indicated that a single LRC was present in an area of 0.072 mm² on average (Table 2) . 

Discussion

Epithelial Cell Movement in the Bulbar Conjunctiva Near the Limbus

The in vivo time-lapse microscopy demonstrated that most of the GFP-positive clusters in the inferior temporal bulbar conjunctiva near the limbus were stationary for at least several weeks and that no directed and persistent movement of GFP clusters took place in this area. These results are in agreement with an anecdotal account of Buck ¹¹ who reported that bulbar conjunctival epithelial cells did not move after a 1-week observation. Histology of wholemount preparations revealed no stripes of GFP-positive cell clusters anywhere in the entire conjunctiva, suggesting that there is little directed cell movement in the conjunctiva. Similar results were reported in a paper by Collinson et al. ¹⁸ Their Figure 3clearly shows patches of lacZ-positive cells without a discernible stripe pattern in the bulbar conjunctiva near the limbus, indicating that this observation is not unique to the GFPU mice. These results suggest that there is little or no directed movement of epithelial cells in the bulbar conjunctiva near the limbus and possibly in other parts of the conjunctiva. If such movement existed, it was too slow to be detected by the methods used in this study. 

These results, however, are contrary to previous reports of conjunctival epithelial cell movement by others who relied on metabolic labeling of DNA and an analysis of tissue cross sections ^{2 8 10} (see the introduction). The reason for this discrepancy is not immediately clear, but it may be because (1) different species exhibit different patterns of cell movement, especially in consideration of various epithelial thicknesses (mouse versus rat and rabbit); (2) kinetics of DNA labeling and dilution or disappearance of labeled DNA is complex, and a temporal change of spatial distribution of labeled cells cannot be directly used to infer cell movement; (3) a small number of cross sections do not represent cell distribution in two dimensions, resulting in a large margin of error; or (4) a combination of any or all of the three. 

Epithelial Stem Cells in the Conjunctiva

Apparently conflicting reports have appeared as to the location of areas enriched in conjunctival stem cells (see the introduction). Our results suggest that epithelial stem cells in the bulbar conjunctiva are distributed uniformly because (1) bulbar conjunctival epithelial cells remained stationary and yet they were mitotically active, indicating that they were capable of self-renewal in situ; and (2) LRCs were distributed uniformly in this area. This conclusion is in agreement with the findings of Pellegrini et al. ³ who demonstrated uniform distribution of cells with a high proliferative capacity, which would include stem cells, in the bulbar conjunctiva in addition to the fornix. Furthermore, our failure to detect directed cell movement in any part of the conjunctiva by histology suggests that the uniform distribution of stem cells may be true in the entire area of the conjunctiva. If there were a large area devoid of stem cells, we would have detected movement of progenitor cells toward such an area, presenting as a GFP stripe. Although the fornix has been reported to contain more stem cells than other areas, ^{4 5 6} it does not seem to be an exclusive area of stem cell distribution. Thus, a high concentration of stem cells in the fornix, if true, may be important in wound healing, but not in conjunctival homeostasis outside the fornix. 

Quantitation of LRC numbers suggested that one LRC is present in approximately 0.072 mm² of bulbar conjunctiva (Table 2) , or approximately 720 basal epithelial cells, assuming the density of 1.0 × 10⁴ cells/mm² in this area (Nagasaki T, Zhao J, unpublished observation, 2004). However, the number of stem cells is probably more than the number of LRCs detected by BrdU labeling, because it is likely that (1) not all stem cells were labeled with BrdU, and (2) some stem cells divided more frequently than others, resulting in the BrdU’s level becoming undetectable with our protocol after a chase period. Thus, the average area occupied by one stem cell would be <0.072 mm² (720 cells) and closer to the average area of GFP clusters, which we calculated, albeit roughly, to be 0.014 mm² (140 cells). In comparison, the estimated number of interfollicular epidermal stem cells ranges from 1 in 10 ^{19 20} to 1 in 10,000 ²¹ basal cells. 

A Boundary of Cell Movement at the Limbus–Conjunctiva Border

We have previously observed that corneal epithelial cells move centripetally at a steady rate of approximately 26 μm/d, starting at the limbal margin. ¹² The present study demonstrated that conjunctival epithelial cells near the limbus are generally immobile. Although we have not studied cell movement within the limbus, our combined results provide direct experimental evidence for a boundary of cell movement at the limbus–conjunctiva border, across which no epithelial cells move in either direction. A question arises as to the molecular and cellular basis of this boundary. On the one hand, the presence of a boundary may not be surprising, considering that the conjunctival and corneal–limbal epithelial cells are thought to be of two separate lineages. ²² On the other hand, the existence of a boundary is difficult to discern, because the mouse limbal epithelium is continuous with both corneal and conjunctival epithelium, and there is no obvious anatomic structure that could assist blocking cell movement. Basement membrane heterogeneity ²³ may be an important mechanism of defining a cellular boundary, as discussed by Wei et al. ²² It is also possible that limbal epithelial stem cell niches provide a structure or a chemical gradient that helps restrict random cell movement. Similarly, stem cell niches of the conjunctiva could play a role in restricting random epithelial cell movement near the limbus. Another possibility is that cell movement is regulated by vascular factors, which would be more or less uniformly distributed in the conjunctiva, due to a capillary network, whereas those in the avascular cornea would form a concentration gradient from the limbus to the central cornea. ²⁴  

 Supported by National Eye Institute Grant EY00431 and Research to Prevent Blindness.

 Submitted for publication March 30, 2004; revised September 27, 2004; accepted October 7, 2004.

 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.

 

Figure 1.

 

Dissection of an eye for preparation of a flat ocular surface wholemount. Skin cuts were made near the margins of both eyelids (A), and the eyeball was enucleated together with the eyelid by dissecting connective tissues around the globe (B). The eyeball was then submerged in saline, and the eyelid was flipped inside out to expose the ocular surface (C–E), yielding a globe with a smooth and continuous ocular surface from the cornea to the mucocutaneous junction (E). (F) A view from the eyelid margin.

Figure 2.

 

Distribution of epithelial GFP in a flat ocular surface wholemount. Ocular surface wholemounts were prepared from the mice of indicated ages, and GFP patterns were imaged with a fluorescence microscope. An entire area of cornea and conjunctiva up to the mucocutaneous junction is shown as an inferior half and a superior half. White areas: GFP-positive cells. In general, goblet cells concentrated in the fornix were the brightest cells. GFP stripes are obvious in the adult corneas, but not in the 4-week-old cornea or the conjunctiva of all ages.

Figure 3.

 

A high-power image of the superior conjunctiva of a 14-week-old mouse. An entire area of conjunctiva from the limbus to the eyelid margin is shown. White areas: represent GFP fluorescence. This specimen did not contain muscle cells, and therefore nearly all GFP fluorescence came from epithelial cells or goblet cells.

Figure 4.

 

A GFP pattern of bulbar conjunctiva in a cross section. Cryosections were prepared from a whole eyeball and stained for nuclei with DAPI. Shown is a section of bulbar conjunctiva near the limbus where in vivo microscopy was performed (see Fig. 6 ). (A) GFP and (B) DAPI staining.

Figure 5.

 

Appearance of goblet cells in the bulbar conjunctiva. Goblet cells were identified by phalloidin staining and also GFP patterns, both of which contained large cytoplasmic vacuoles, located in the bulbar conjunctiva away from the limbus (A–C), but not near the limbus (D–F), which is the area in which in vivo time-lapse recordings were obtained (see Fig. 6 ). Shown are DAPI nuclear staining (A, D), GFP distribution (B, E), and phalloidin-Alexa 546 staining (C, F).

Figure 6.

 

A time-lapse sequence of GFP-positive cell clusters in the bulbar conjunctiva near the limbus. Overlapping microscopic areas were captured and later patched together to generate an image of bulbar conjunctiva near the limbus. Images were recorded at least once a week. The time points are the ages of the animal. White areas: GFP-positive cells. (Arrowheads, panels 17–21.3 weeks) a group of GFP clusters that maintained general shape and positions. These clusters changed positions drastically between 21.3 and 21.6 weeks. (Arrows, panels 21.6–52 weeks) a group of GFP clusters in the same area that remained mostly stationary for the duration of the observation.

Table 1.

 

A List of Time-Lapse Sequences That Were Analyzed^*

Mouse ID	Age at the Start (wk)	Length of Observation (wk)	Number of GFP Clusters, ‡
						Analyzed Total	Directed Movement for 3 Weeks	Stationary for 8 Weeks
1	11	15	13	0	10
2	12	20	9	0	7
3	13	24	5	0	4
4	13	11	8	0	4
5, †	14	38	9	0	7
6	14	14	9	0	6
7	16	9	9	0	5
8	16	20	9	0	6
9	17	14	12	0	9
10	24	29	7	0	5
Total			90	0	63

*  Only those recordings that continued for more than 8 weeks are included.

†  Selected time-lapse images between 17 and 52 weeks are shown in Figure 6 .

‡  The number of GFP clusters was determined from time-lapse sequences.

Figure 7.

 

DNA synthesis in the bulbar conjunctiva near the limbus. DNA was metabolically labeled with BrdU in vivo, and BrdU-labeled nuclei were detected by immunohistochemistry. Triple fluorescence images of the same area are shown for GFP, BrdU, and DAPI.

Figure 8.

 

Distribution of LRCs in the superior bulbar conjunctiva of an 8-week-old mouse. A wholemount preparation of the entire superior conjunctiva is shown as a background image with reduced contrast, with the eyelid margin on top. To obtain flatmounted specimens, slits were made, and the cornea including the limbus was removed from the conjunctiva. LRCs were identified by BrdU immunohistochemistry after a 7-week observation, and are plotted as individual dots. Cells were counted in an area up to 0.8 mm from the limbus, demarcated by a solid line. Shown is a representative of six superior and six inferior conjunctival preparations, all of which showed a similar pattern of LRC distribution.

Table 2.

 

Quantitation of LRCs in the Bulbar Conjunctiva

	Total Number of LRCs	Measured Area (mm2)	Average Area per Single LRC (mm2)
Superior conjunctiva	48.8 ± 15.4	3.32 ± 0.28	0.072 ± 0.017
Inferior conjunctiva	47.5 ± 17.2	3.24 ± 0.41	0.073 ± 0.017
Superior + Inferior	96.3 ± 31.6	6.56 ± 0.42	0.072 ± 0.016

 * All LRCs (BrdU positive cells) were counted in the bulbar conjunctiva (n = 6) up to 0.8 mm from the limbus (i.e., demarcated with a solid line in Figure 8 ). The data are the mean ± SD of six eyes from three mice. Note that the measured area is dependent on how an elastic piece of wholemount conjunctiva was mounted on a slide, and therefore it may not represent the surface area in a living eye.