Article Text

Download PDFPDF

Original article
In vivo confocal microscopic features of normal limbus
  1. Ammar Miri1,2,
  2. Muhamed Al-Aqaba1,
  3. Ahmad Muneer Otri1,2,
  4. Usama Fares1,
  5. Dalia G Said1,3,
  6. Lana Akram Faraj1,
  7. Harminder S Dua1
  1. 1Division of Ophthalmology and visual sciences, University of Nottingham, Nottingham, England, UK
  2. 2Department of Ophthalmology, Aleppo University, Aleppo, Syria
  3. 3Research Institute of Ophthalmology, Cairo, Egypt
  1. Correspondence to Dr Harminder S Dua, Division of Ophthalmology and Visual Sciences, B Floor, Eye ENT Centre, Queens Medical Centre, University Hospital, Derby Road, Nottingham NG7 2UH, England, UK; harminder.dua{at}nottingham.ac.uk

Abstract

Aim To describe in vivo confocal microscopy (IVCM) features of the limbus in normal eyes as related to the palisades of Vogt's.

Methods 46 eyes of 29 consecutive volunteers were recruited in this observational study. A detailed examination by IVCM was performed in addition to a routine slit-lamp biomicroscopy. Size and density of corneal and limbal epithelial cells were measured and statistically analysed using SPSS version 8.0 software.

Results Anatomical and morphological features were noted between corneal and limbal cells. Size and density differences reached to significant levels (p<0.05). Different shapes of palisades of Vogt have been described clearly by confocal microscope. Cell-like structures were observed in the peripheral end of the palisades which might represent limbal stem cell crypts.

Conclusions Laser IVCM can be used to establish the features of the normal limbus. The identified features demonstrate quantitative changes in the basal epithelium between the limbus and the central cornea and morphological differences between pigmented or non-pigmented studied subjects. Further studies should be performed to correlate with histology the possible crypts which were observed in this study.

  • Palisades of Vogt
  • confocal microscope
  • cornea
  • iris
  • microbiology
  • infection
  • tears
  • pathology
  • angiogenesis
  • anterior chamber
  • conjunctiva
  • clinical trial
  • degeneration
  • treatment surgery
  • inflammation
  • medical education
  • wound healing
  • eye (tissue) banking
View Full Text

Statistics from Altmetric.com

Introduction

The limbus is the junction between the cornea and the sclera and is divided into the corneal limbus (CL), and the more posteriorly located scleral limbus (SL).1 2 Clinical and laboratory evidence suggests that the niche for corneal epithelial stem cells is located in the corneoscleral limbus.1 3–5 This makes the limbus a site of great interest to both clinicians and researchers. Of particular interest are the palisades of Vogt (POV), which are a series of fibrovascular palisade-like structures found predominantly along the superior and inferior limbus, but seem to be more regular and prominent at the lower limbus.6 These ridges are covered with 2–3 layers of epithelium and alternate with the epithelial rete pegs, which are made of 10–15 layers of epithelial cells.1 6 7 The inter-palisadal epithelial rete pegs may play an important role in providing the microenvironment or niche for limbal stem cells.6 Melanin pigment is believed to afford protection to stem cells from the deleterious effect of exposure to ambient ultraviolet radiation. The amount of pigment is directly related to the skin colour and in dark(er) individuals details of the pigmented palisades are clearly visible even on slit-lamp examination. Recently, solid cords of cells extending from the posterior end of the limbal epithelial rete pegs have been identified. These anatomical structures were termed the limbal epithelial crypts (LECs).3 There is evidence to suggest that these LEC are putative stem cell (SC) niches.5 8

The introduction of the in vivo confocal microscope (IVCM) provides high resolution images of living tissues which enables us to study the micro anatomical features of transparent, translucent and semiopaque structures on the ocular surface. Recently, IVCM has been used to obtain high quality images of the limbus including the POV. In this study, we will describe the IVCM features of normal limbus with emphasis on the examination of pigmented palisades.

Methods

A total 46 eyes of 29 healthy subjects, 17 were males and 12 females, were studied after obtaining the consent of the subjects. Ethics approval for this study was obtained from the local Ethics Committee (No. 06/Q203/46). Exclusion criteria were absence of the palisades on the slit-lamp biomicroscopy, previous ocular surgery, <18-years of age, any current or past history of ocular surface disease including dry eyes, current or previous use of any topical eye medication for more than 1 week duration, previous eye injury, contact lens use and systemic disease with eye involvement.

The eyes studied were evaluated according to age and sex of the subjects and pigmentation of the palisades. Eyes were divided into two groups: Group A included subjects whose age was <40 years and group B included subjects of greater than or equal to 40 years. With regard to pigmentation, the subjects were divided into three groups: non-pigmented subjects (European Caucasian), moderate pigmented subjects (Asian and Middle East) and hyperpigmented subjects (African and black Indian). A detailed history and clinical examination with slit-lamp biomicroscopy was carried out to exclude any existing or previous eye disease or surgery. The palisades were examined and 40× slit-lamp photomicrographs of the palisades were taken.

A detailed IVCM examination of the central cornea (CC) and upper and lower limbus was performed using the Heidelberg Retina Tomograph II Rostock Corneal Module (Heidelberg Engineering GmBH, Dossenheim, Germany).9

Mean epithelial cell diameter and density of the CC, superior and inferior CL and superior and inferior SL were measured and averaged from three frames of each layer examined. The average value for basal and wing cells at each site was also determined. Cell density was calculated using the original software of the confocal microscope whereas the cell diameters were measured using ImageJ software (ImageJ V.1.31, Maryland, USA). Data were statistically analysed through SPSS program statistical package V.8.0 (SPSS Inc.). The normality of the data was tested prior to analysis. The t test was used for normally distributed data and the Wilcoxon signed rank test was used for data that were not normally distributed.

Results

The patients' ages ranged from 21 to 62 years with a mean of 44±12 years. Group A included 23 eyes of 14 patients (6 female and 8 male subjects) with a mean age of 33±5 (range 21–39 years). The subjects comprised four white Europeans, four Asians or Indians and six Africans. Group B included 23 eyes of 15 patients (6 female and 9 male subjects) with a mean age of 55.4±6 (range 42–62 years). The subjects comprised five white Europeans, six Asians or Indians and four Africans. On slit-lamp examination POV were visible as radially arranged linear or branching structures. Pigmented palisades were observed in 10 eyes in group A and six eyes in group B, which were all in subjects of African or Indian origin, whereas the palisades were non-pigmented in 30 eyes (13 eyes in group A and 17 eyes in group B) of white European, Asian or Indian subjects (figure 1).

Figure 1

Diffuse slit-lamp images of the cornea. (A) Inferior limbus of a non-pigmented subject: the palisades of Vogt (POV) were not pigmented. (B) Same eye with higher magnification of the inferior limbus: the POV were visible as linear structures in the limbus (arrows). (C) Inferior limbus of a hyperpigmented subject showing pigmented POV. (D) Same eye with a higher magnification of the inferior limbus: the POV were clearly visible (arrows). The projections of the POV could be observed towards the corneal side (arrow heads).

IVCM

IVCM of CC

Scanning of the CC revealed normal features in all eyes which have been described previously.10 11 Briefly, superficial cells presented bright cytoplasm and bright nuclei. The wing cells were smaller than the superficial cells with dark cytoplasm and well-defined bright borders (figure 2A). The basal cell layer too appeared dark with bright borders but cells were even smaller in size (figure 2B). Unlike the superficial cells, nuclei in the wing cells and the basal cells were not seen by IVCM. Cell morphometry data are given in table 1.

Figure 2

In vivo confocal microscopy images of the central cornea (CC) and corneal limbus (CL). (A) Normal wing cells of the CC appeared with a dark cytoplasm with well defined bright borders and no visible nuclei. (B) Basal cells of the CC were smaller than the wing cells with no visible nuclei. (C) Hyper-reflective cells of dendritic morphology were observed among the basal cells of the CL in 19 eyes. (D) Curvilinear columns or streaks, interrupting the regular mosaic of basal cells, were observed towards the corneal side of the CL. (E) CL of a moderately pigmented subject: hyper-reflective cells were observed among the normal wing and basal corneal epithelial cells. (F) CL of a hyperpigmented subject: hyper-reflective cells were more prominent and occurred with higher frequency than in the moderately pigmented subjects.

Table 1

Morphometry data of the central corneal epithelium

IVCM of CL

Epithelial cells of CL were similar in morphology to those visualised in the CC. Hyper-reflective cells of dendritic morphology were also observed among the basal cells in 19 eyes (figure 2C). The mean density of these cells was 267±119. Dark, curvilinear columns or streaks, interrupting the regular mosaic of basal cells, were observed towards the corneal side of the CL (figure 2D). In moderately and highly pigmented subjects, hyper-reflective cells were observed among the normal wing and basal corneal epithelial cells (figure 2E,F).

The mean cell diameters and densities for wing and basal cells in the superior and inferior CL are given in table 2. There was a significant difference in the cell diameter and density of the basal cells in group A compared with group B (p value <0.03 and 0.04, respectively), that is, younger subjects had smaller cells and greater density.

Table 2

Morphometry data of the corneal limbus epithelium

IVCM of SL

POV were visible in the superior and inferior SL of all eyes. They presented as hyper-reflective, double contoured, usually parallel lines. These alternated with islands of epithelial cells which corresponded to the inter-palisade rete pegs (figure 3A). Some POV presented a circular or oval shape (figure 3B). The adjacent palisades were often connected to each other (figure 3B,C).

Figure 3

In vivo confocal microscopy images of the scleral limbus. (A) An eye of a non-pigmented subject: the palisades of Vogt (POV) were observed as hyper-reflective, double-contoured linear structures (arrows). These alternated with islands of epithelial cells which corresponded with rete pegs (arrow heads). (B) An eye of hyperpigmented subject: circular or oval shape POV were observed. The POV were connected to each other. Numerous hyper-reflective cells are observed in the rete pegs. (C) An eye of hyperpigmented subject showed connections between the palisades (arrows) and hyper-reflective structures at the level of the basal cells of the rete pegs (arrow heads). (D) An eye of a non-pigmented subject: it was difficult to distinguish the borders of the basal cells of POV (arrows). (E) An eye related to a moderately pigmented subject: the borders of the basal cells of POV were better defined in moderately pigmented than in non-pigmented subjects (arrow). The basal cells of the POV created a curved demarcation line just central to the conjunctival cells and at the termination of the palisades. (F) An eye of a hyperpigmented subject: numerous hyper-reflective cells were observed in the rete pegs and the basal cells of the POV. Towards the corneal side the projections of the POV were visible (arrow). At places, it was difficult to distinguish the cell borders within an excessive hyper-reflective area (arrow heads).

The mean width of the palisades was 22±7.6 μm. In non-pigmented subjects (14 eyes of 9 subjects) the basal cells of the palisades (at the summit or along the sides) were brighter and had ill-defined margins compared with cells in the rete pegs. The cells in the rete pegs were dark with well-defined borders. The nuclei were not visible in any of the cells (figure 3A,D). In moderately pigmented subjects (16 eyes of 10 subjects), hyper-reflective cells were seen among the basal cells of the POV (figure 3E). In hyperpigmented subjects (16 eyes of 10 subjects) there were numerous such hyper-reflective cells, which were also visualised within the rete pegs (figure 3B,F).

The borders of the basal cells of POV were better defined in moderately pigmented than in non-pigmented subjects. Their definition allowed morphometry in five eyes superiorly and inferiorly (three eyes in group A and two eyes in B) (figure 3E). The basal cell borders were even better defined in hyperpigmented subjects (figure 3F). In some eyes of moderately pigmented and hyperpigmented subjects the basal cell clearly delineated the outline of the palisade presenting as curved demarcation lines just central to the conjunctival cells and at the termination of the palisades (figure 3E). The palisades were wider at the periphery (conjunctival end) and gradually tapered centrally. At places, the excessive reflectivity of the basal cells made it difficult to distinguish individual cells within a hyper-reflective area (figure 3F).

Towards the central (corneal) end of the palisades, small rosettes of hyper-reflective cells surrounding dark central areas were seen in all subjects. These too were best defined in hyperpigmented individuals and less distinct in non-pigmented subjects (figures 3D–F and 5).

The reflectivity of the POV and the inter-palisade rete pegs was found to differ according to depth. It was noticed that the palisades were hyper-reflective and the cells in the rete pegs were darker superficially (mean depth of 70 μm) (figure 4A). However, deeper in the SL, the pattern of reflectivity reversed. The area of the rete pegs became hyper-reflective once the plane of scanning had gone beyond the basal cells and alternated with darker areas which corresponded to the palisades above (mean depth of 120 μm range from 95 μm to 210 μm) (figure 4B,C). Also, we observed interconnections between the deep stroma of adjacent palisades, presenting as hypo-reflective transverse bands (figure 4D–F).

Figure 4

In vivo confocal microscopy images of the scleral limbus (SL). (A) Hyper-reflective palisades of Vogt (POV) (arrows) and darker cells in the rete pegs (arrow heads) were observed superficially at the SL. (B) The same eye, at a deeper level, beyond the basal cells at the SL: the area of the rete pegs became hyper-reflective (arrow heads) and alternated with darker areas corresponding to the palisades (arrows). (C) A hyperpigmented eye showing the same pattern at the SL with more hyper-reflectivity at the deep area beyond the rete pegs (arrow heads). (D–F) Volumetric images of the stroma beneath the POV showing interconnections between the deep stroma of adjacent palisades, presenting as hypo-reflective transverse bands (arrows). (G) SL in a hyperpigmented subject showing the POV and basal cells of the rete pegs. (H) Same eye at a deeper level: clusters of cells with hyper-reflective basal cell morphology were observed (arrow). (I) The same finding (as H) in another eye of a hyper-reflective subject was seen at a deeper plane in the SL (arrow).

Figure 5

In one eye of a moderately pigmented subject, multiple overlapping images of the superior limbus were obtained and organised into a wide-field montage. Palisades of Vogt with projections towards the corneal side can be seen. Dark curvilinear columns or streaks, interrupting the regular mosaic of basal cells, were observed towards the corneal side of the corneal limbus adjacent to the central corneal stroma (arrows).

In nine eyes of hyperpigmented subjects (five in group A and four in group B), clusters of cells with the hyper-reflective basal cell morphology described above were seen in the deep plane, surrounded by collagen of the limbal stroma. These were present at the conjunctival termination of the palisades at a mean depth of 124 μm (range from 111 to 142) (figure 4G–I). These cells were small in size with a mean diameter of 10.1 μm and 10.3 μm in groups A and B, respectively, ranging from 9 μm to 11 μm and a mean density of 7815±723 cells/mm2 and 7759±783 cells/mm2 in groups A and B, respectively.

Figure 5 shows a montage of IVCM images of the superior limbus in a moderately pigmented subject.

The mean cell diameters and densities for middle and basal cells of the rete pegs in the superior and inferior SL are given in table 3. A significant difference was found between group A and group B in relation to the diameter and density of cells located in the middle of the rete pegs (depth 40–50 μm) (p value <0.05) and also between the basal cells of the rete pegs of the two groups (p value <0.04 and 0.004, respectively). The younger the subject the smaller was the cell diameter and the greater was the density.

Table 3

Morphometry data of the scleral limbus epithelium

The cells arranged along the sides of the POV presented their long axis to the scanner. Thus, it was possible to measure cell height (length) and diameter (width) which was not possible with cells in the rete pegs, which presented only their diameter (width) to the scanner. The mean height (length) of the basal cells of the POV when the borders were visible (13 eyes in group A and 6 eyes in B) was 12.3±1.9 μm and the mean diameter (width) was 10.2±0.9 μm in group A. In group B, the same dimensions were 12.4±1.5 μm and 10.1±1.1 μm, respectively.

Analysis

No significant variations of cells diameter and density were recorded between male and female subjects in both groups. A significant difference was noted for the diameter and density of basal cells between the three studied areas (CC, CL and SL) in eyes of the same group, that is, the closer to the limbus the smaller the diameter of the basal cells and greater the density. p Values were <0.001 for all variations. Also, a similar significant difference was found between the diameter and density of the corneal wing cells (central or peripheral) and the middle cell layers of the rete pegs in the same group (p values were <0.001 for the diameter and <0.001 for the density).

No significant difference was noted between the basal cell diameter in the rete pegs and the width of the basal cells of the palisades (p values were 0.4 in group A and 0.3 in group B). Also, no significant difference was recorded for the diameter and density between the basal cells of the rete pegs and the cells which were visible in the deep stroma at the conjunctival side of the POV (p values were 0.5 and 0.4 respectively for group A and 0.3 and 0.4 respectively for group B).

Discussion

Scanning of the limbus by confocal microscope enables the examiner to approach the limbus in its physiological state; furthermore, it has the advantage of reducing the light scatter from the sclera which was a challenge with the white light microscope.1 It has been reported that the palisades are not visualised in 10%–20% of the population, particularly in non-pigmented subjects and in older subjects.1 12 In this study, the absence of the POV on slit-lamp biomicroscopy was one of the exclusion criteria and hence we were able to examine the POV by IVCM in all eyes included.

The hyper-reflective cells at the limbus, which were more prominent with increased pigmentation, suggest that the melanocytes render these cells more visible during confocal microscopy. The presence of melanocytes has been previously confirmed by histological and electron microscopic studies.12 13 It was suggested by Cotsarelis et al that the pigment plays a vital role in protecting the limbal stem cells from solar damage.14 In POV of non-pigmented subjects the visualisation of the palisade basal cells was poor.

In this study, a cord of cells extending into the conjunctival stroma just peripheral to the POV were seen in nine eyes of highly pigment subjects. These structures probably represented the LECs reported.3 5 Such extensions have been reported in only one previous study on IVCM of the limbus.15

In our study, these cords of cells were observed in the stroma of the conjunctival end of the POV at a mean depth of 124 μm (range from 111 to 142). These cells were significantly smaller in diameter than the basal cells of the CL and CC indicating a different cell phenotype. Cells of the LEC have been reported to be different in morphology and immunophenotype compared with the other cells.5 Furthermore, there was no significant difference in the diameter or the density of the basal cells of the rete pegs and that measured for the cord cells, which suggests a similar cell phenotype.

Hyper-reflective cells of dendritic morphology observed in the CL have been previously reported16 and probably represent Langerhans (dendritic) cells. These are known to migrate to the CC following inflammation.17 The presence of a border of dark and hyper-reflective cells in the limbus represents the transitional zone between the corneal phenotype and the conjunctival phenotype.16 18

Unlike the central basal epithelial cells, the diameter and density of cells in the limbus change with increasing age whereby the mean diameter increases with a corresponding decrease in cell density.1 15 Our findings confirmed this observation.

Moreover, this study shows significant differences in diameters and densities of the basal epithelial cell among CC, CL and SL. The diameters were the greatest in the CC whereas the density was the greatest in the rete pegs. Kobayashi and Sugiyama reported that cell sizes of the central corneal basal epithelial cells were significantly larger than those ‘beneath the POV’.19 Similar results were found by Romano et al although they reported a larger cell size both in cornea and limbus. This may in part be related to the different confocal equipment they used.20

Our findings regarding the cell diameters and densities did not concur with the findings of Patel et al who reported the highest density and the smallest diameters of the basal cells in the CL.1 They could not visualise the basal cells in the palisades because of their poorly defined borders and hence could not calculate their density. They therefore used the epithelial cells within the rete pegs as a surrogate for limbal palisade cell density but no details of the depth at which the measurement was performed are provided. In our study, we refer to the basal cells of the limbus as the deepest visible cells in the rete pegs. IVCM could not measure the area of the footprint of the basal cells along the palisade ridges as these cells lay horizontally in relation to the objective of the confocal microscope. This only allowed measurement of the width and height of these cells. In our pigmented palisades, we were however able to observe the footprint area of the basal cells of the rete pegs and hence calculate the density of these cells.

One interesting and novel observation that we made in this study is that the stromal architecture of the palisades was maintained beyond the plane of the basal cells of the rete pegs. When the IVCM was focused at planes deeper than the basal cells of the rete pegs, alternating rows of bright and dark images were still visible as illustrated in figure 4B,C, with the bright rows representing the overlying ‘rete peg area’ and the dark rows representing the overlying ‘palisade area’. This reflectivity was converse of that seen in the planes anterior to the basal cells where the cellular rete pegs are hypo-reflective and the palisades are hyper-reflective. This pattern was best seen in non-pigmented subjects and less so in the pigmented subjects due to the increased general level of reflectivity in the latter group. Furthermore, we observed interconnections between the deep stroma of adjacent palisades, presenting as hypo-reflective transverse bands (illustrated in figure 4D–F), suggesting that there exists a ‘foundation’ to the palisades in the deep stroma, upon which stands the stromal support, which holds the cells of the palisades and rete pegs.

Another interesting and novel finding was related to the dark, curvilinear columns or streaks, interrupting the regular mosaic of basal cells: these were observed towards the corneal side of the CL. These most probably represent folds or undulations in the basement membrane and underlying stroma, before these even out as a smooth layer over the Bowman's zone. The dark lines (figures 2D and 5) would correspond to the folds, in which no cells would be present. This appearance is also seen in the images presented in another publication1 but the authors did not comment on these. Furthermore, similar dark streaks have been proposed as a feature of corneal basement membrane dystrophy21 where they correspond to basement membrane folds, supporting the suggestion that they are caused by folds of the connective tissue under the basal cells.

Our study is important because, by specifically scanning pigmented palisades, we were able to attain better defined images and introduce more accuracy in the morphometric data compared with previous publications.1 19 Furthermore, we were able to confirm the presence of cords of cells which extended into the conjunctival stroma just peripheral to the POV, which correspond to the histologically described LECs.3 5 Three other novel aspects of our study are the demonstration of the characteristics of the basement membrane and underlying stroma of the limbus including palisades and rete pegs, which showed that the stromal architecture is maintained beyond the plane of the basal cells of the rete pegs. Second, we have demonstrated the presence of folds or undulations in the basement membrane and underlying stroma extending from the palisades to the CL and finally we were able to confirm that there is no difference between the superior and inferior limbus with regard to cell size and density.

References

View Abstract

Footnotes

  • Linked article 300551.

  • Competing interests None.

  • Patient consent Obtained.

  • Ethics approval Ethics approval was provided by Nottingham research ethics committee, No. 06/Q203/46.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles