Noncontact Specular Microscopy of Human Lens Epithelium

Abstract

purpose. To obtain in vivo specular images of human lens epithelial cells (LECs) from persons with or without age-related cataract (ARC); to identify features that describe individual aspects of these complex images; to develop feature scales to quantify the severity of each feature; and to study the association of these features with LEC count, age, Lens Opacity Classification System III (LOCS III) classifications and microscopic features of lens epithelium in ARC.

methods. One hundred fifty-two individuals underwent ophthalmic examinations and LOCS III cataract classifications. Specular images of lenses were captured using a modified noncontact corneal specular microscope (SML-2; Konan, Hyogo, Japan). Enhanced images were graded in a masked fashion, and the presence or absence and severity of each of four features in the specular image (“columnar organization,” “linear furrows,” “puffy clouds,” and “black holes”) was graded on a four-step scale. The generalized linear model with intraclass correlation was used to ascertain the statistical significance of associations between age, sex, LOCS III grade, cell count, and feature grade. Capsulorrhexis specimens from 29 patients were studied with correlative light and electron microscopy.

results. LEC density declined with age and was inversely correlated with the scalar grade for puffy clouds and for the size and number of black holes. The scalar grade for columnar organization was inversely associated with the severity of posterior subcapsular and nuclear cataracts, which was the only feature associated with the LOCS III grade of ARC. No statistically significant associations were found between average cell count and LOCS III grade.

conclusions. With the use of the corneal specular microscope excellent in vivo specular images of the LECs were obtained, the features in these images that correlated well with microscopic findings were classified, and cell density in vivo was estimated.

The contribution of LEC injury to cataractogenesis has been studied in different models. ^{1 2} Corticosteroids and galactose slow epithelial replication, alter epithelial cytology, and hasten senescence. Eventually, such stresses cause lens opacification. ^{3 4} These studies have helped focus interest on the epithelial cells of the aging human lens and their role in age-related cataractogenesis. ^{5 6} There is controversy about the effects of age on average LEC density. Guggenmoos–Holzmann et al. ⁷ noted a decline with age of 7.8 cells/mm² per year. Karim et al. ⁸ and Fagerholm et al. ⁹ found no correlation between cell count and age. Karim et al. showed that the LEC density was lower in cortical, posterior subcapsular, mixed posterior subcapsular /cortical, and posterior subcapsular /nuclear cataracts. They showed also that the LEC mitotic index declined with age both in cataractous and noncataractous lenses. Perry et al. ⁵ and Konofsky et al. ¹⁰ found no significant change in LEC density with early cataract but noted that it was markedly reduced in advanced cataract. An increase in LEC size with age has been documented by multiple observers. ^{4 5 6}  

LEC density in women is higher than in men in both clear and cataractous lenses. ^{7 10} Cataractous lenses contain a mixture of normal and abnormal LECs. ¹¹ The abnormal features include altered hexagonal cellular arrays, extensive intercellular uncoupling and decreased cytologic activity as inferred from changes in morphology and amounts of rough endoplasmic reticulum, Golgi apparatus, and mitochondria. ¹¹ Vacuolization ¹² and increased granularity of cytoplasm are seen in light micrographs of older lenses, ⁵ and enlarged extracellular lacunae are seen in electron micrographs of cataractous lenses. ¹¹  

In advanced cataracts, metaplastic changes are evident; mitotically active cells are superimposed on the central epithelial zone, the nuclear size is altered, and degenerated nuclei are more evident. ^{12 13} LEC size and pleomorphism also increase with the severity of cataract. ^{3 6}  

In 1951, Francois and Rabaey, ¹⁴ used phase-contrast microscopy to study the LEC layer in vitro “in its natural living state,” unaltered by fixation and staining. They found two distinct cell types: “pale polyhedral” and “dark star shaped.” They also recognized that increasing intracellular vacuolization preceded cell death in bovine LECs. 

Considerable work has been done to image LECs in vivo. ^{15 16 17 18 19 20} The first study cited was completed in 1975 by Laing et al. In 1981, Laing and Bursell ^{15 16} modified a corneal specular microscope to study the corneal endothelium and measured the diameter of the LEC and the thickness of the lens fiber. This instrument used a water-immersion objective coupled to a suction contact lens, and it captured high-magnification (×50) images of the anterior and posterior lens surface. Further refinement of the technique of specular photography of the lens epithelium yielded images of increasing clarity and detail. ^{17 18 19 20} Using a noncontact method, Bron and Matsuda ¹⁷ demonstrated three distinct specular zones in the anterior human lens: coarse shagreen, epithelial pattern, and lens fiber pattern. They were shown to represent the capsule, the epithelial cell layer, and the lens fibers, respectively. Oak et al. ¹⁸ using a contact specular microscope, identified surface folds and attributed them to lens fibers. They also noted bubble-like structures on the anterior surface of cataractous lenses and suggested that they represented swollen epithelial cells. In 1987, Brown and Bron ¹⁹ using noncontact specular reflex photography were able to assess the size and density of human LECs. In their study of 50 eyes of a 100 subjects between the ages of 11 and 75 years, they showed that the in vivo diameter of the LEC ranged from 9 to 17 μm and that LEC pleomorphism existed in each individual. No associations between LEC size and age or the presence of diabetes mellitus were found. Masters et al. ²¹ reported success with confocal light microscopy to image human LECs in vitro and in an in situ lens from an ex vivo rabbit eye. ²² However, no in vivo studies using this imaging modality have been reported. 

In our study, we used noncontact specular microscopy with image enhancement and light and transmission electron microscopy to clearly reveal several age and cataract-related features in the complex images of the human lens epithelium. The Specular Microscope (for) Lens (SML-2; Fig. 1 ) a modified corneal specular microscope, (SP8000, Konan; Hyogo, Japan) ²³ easily captured digital images of LECs in both clear and cataractous lenses. It focused on the third refractive index change from the anterior surface of the eye. The digital images were stored as tagged image file format (TIFF) files by interfacing the corneal specular microscope to a computer using a frame grabber board (model DT55; Data Translation, Marlborough, MA). The original image was enhanced to better define the cellular boundaries and morphologic features of the LECs. We developed a system to grade these features, to count the number of LECs per square millimeter, and to study their association with age, gender, and age-related cataract (ARC), and we correlated the structure of LECs in capsulorrhexis specimens with the features of specular images. 

Methods

Study Population

One hundred fifty-two individuals of ages ranging from 20 to 89 years were recruited for the study from the Division of Ophthalmology at the Brigham and Women’s Hospital, according to the hospital’s Human Studies Committee approved protocol. The study population consisted of sixty-one (40%) men and 91 (59.9%) women. Two hundred seventy-four eyes were studied. Of these, 254 were available for cortical and posterior subcapsular studies, and 252 were available for nuclear opalescence and nuclear color studies (see Table 1 ). One hundred and twenty-two subjects had both eyes evaluated, 17 (11.2%) had examination of only the right eye, and 13 (8.6%) of only the left eye. 

Procedure

The purpose of the study, the procedure, including possible risks, was explained in detail to each study participant, in accordance with the Helsinki Declaration. All 152 subjects elected to participate in the study. After providing informed consent, each subject underwent a standard ophthalmic evaluation. This included an assessment of visual acuity, applanation tonometry, slit lamp biomicroscopy of the anterior segment, pupillary dilation using 2.5% phenylephrine and 1% tropicamide, and Lens Opacities Classifications III (LOCSΙΙΙ ) ²⁴ classification of ARC. Nidek (Nidek, Tokyo, Japan) EAS 100 Scheimpflüg and retroillumination photographs were taken and used for LOCS III grading. The LOCS ΙΙΙ grades ranged from 0.1 to 4.0 (cortical; C), 0.1 to 4.9 (posterior subcapsular; P), 0.1 to 5.8 (nuclear opalescence; NO), and 0.1 to 6.8 (nuclear color; NC). The number of eyes in each range of LOCS ΙΙΙ grading scale for each class of cataract is presented in Table 1 . The specular images stored as TIFF files were processed with image analysis software (Global Laboratory Image Software; Data Translation, Marlborough, MA). Enhancement of the images better defined the borders of individual cells. 

The cell density was determined by a grid method in which a square area of known size was superimposed on the enhanced specular image, and the cells were manually counted. Cells overlapping two borders of the square (for example, upper and left) were counted, and those overlapping the other two sides (lower and right) were ignored. An average of three readings was taken for each eye. 

To develop a feature set, all enhanced specular images were examined as a group and common features identified. Care was taken to ensure that only those features that could be graded reproducibly were included. Four such features were identified, and scales representing minimal, moderate, and advanced stages of each feature were constructed. Three images (as 5 × 7-in. black-and-white prints) were selected from the entire set to represent the boundaries between minimal and moderate, moderate and advanced, and advanced and very advanced. Each scale consisted of four steps or grades, with grade 1 indicating minimal involvement and grade 4 indicating maximal involvement. 

The 274 specular images were then masked to the individual’s age, sex, and LOCS ΙΙΙ classification of ARC. The severity of each feature in each specular image was graded independently by two experienced observers. The two sets of grades were compared, there was excellent agreement between graders, and in the few cases in which the scores differed, the differences were not more than one step. In these cases a consensus grade was selected. The data thus obtained were analyzed using the generalized linear model with intraclass correlation ²⁵ to identify statistically significant associations between feature severity, age, sex, LOCS III classification and LEC density. 

To ascertain the anatomic change in the lens epithelium responsible for these specular image features, we identified 29 subjects in the study population who had been scheduled for cataract surgery. A protocol for the collection of anterior capsules of these subjects at the time of surgery was submitted and approved by the Human Studies Committee of the Brigham and Women’s Hospital. Informed consent was obtained before surgery for the intraoperative procurement of central curvilinear capsulorrhexis specimens. Immediately after completing the capsulorrhexis, the surgeon, (LTC) irrigated the specimen with balanced salt solution (BSS Plus, Alcon, Fort Worth, TX) to remove blood and viscoelastic and placed it directly into 2.5% glutaraldehyde in 0.12 M Na cacodylate buffer (pH 7.2), where it remained for 1 hour at room temperature. It was then gently removed, rinsed twice in 0.2 M Na cacodylate buffer (pH 7.4), immediately stored in the same buffer, and sent to coauthor JRK for light and electron microscopic studies. 

Light and Transmission Electron Microscopy

After a 1-hour rinse in cacodylate buffer, specimens were further preserved in 1% osmium tetroxide for an additional hour at 4°C. After a 1-hour rinse in cacodylate buffer, specimens were dehydrated through a graded series of ethanols, 100% first, and then through a graded series of ethanol-propylene oxide to pure propylene oxide. Specimens were then infiltrated and embedded in Epon 812 resin. For light microscopy, thick (1–2 μm) sections were cut with glass knives on a microtome (model 8800 Ultratome; LKB, Bromma, Sweden), mounted on glass slides (Premount; Fisher Scientific, Fairlawn, NJ), and stained with a dilute mixture of methylene blue and azure II. A photomicroscope (Vanox AHBS3; Olympus America, Melville, NY) was used to examine the specimens. 

For transmission electron microscopy , thin (60-nm) sections were cut with diatome diamond knives (Diatome–US, Fort Washington, PA), mounted on uncoated 200-mesh copper grids. The sections were stained with 1% uranyl acetate and lead citrate, and examined (model 1200EX; JEOL USA, Peabody, MA) at 60 kV. 

Results

The four features identified in the specular images were described as “linear furrows,” “columnar organization,” “puffy clouds,” and “black holes.” The feature black holes could be subclassified according to the average size of the black holes and the number of black holes per square millimeter (see Figs. 2A 2B 2C 2D 2E ). 

In Figure 2A linear furrows are visible. These are black spindle- or furrow-shaped features in the left image that occupy most of the lower right quadrant of the image. In the left-hand image, they run from lower left to upper right through the entire image as irregular black columns separated by whitish reticular patterns. This feature resembles the furrows in a plowed field. 

In Figure 2B , almost as if through the reticular pattern of the epithelial cells, vertical lines or columns are visible. This is the feature we called columnar organization. These columns are seen faintly in the image on the left, whereas they are very evident in the image on the right. 

Figure 2C shows the feature puffy clouds, which are whitish cloudlike areas. In the right-hand image they appear almost to be floating above the reticulated epithelial pattern below. This feature is only slightly evident in the image on the left. 

In Figure 2D we see the variation in the size of the feature black holes. The image on the left illustrates small holes, whereas that on the right shows larger holes. The variation in the number of black holes is shown in Figure 2E . The image on the left contains only a few (<20) black holes in comparison with the image on the right, which contains many more. 

It is evident in the figure that there are other features in these complex images (e.g., pleomorphism of LECs), but these could not be measured or graded reproducibly. The four principal features could be identified easily and graded reproducibly and therefore offered an opportunity to study the anatomic bases for the specular changes. 

The grade of the feature columnar organization was highly significantly associated with the LOCS ΙΙΙ grades of P, NO, and NC (see Table 3 ). With increasing severity of P or NO, there was a reduction in the columnar organization of the LECs. With increasing brunescence (NC), there was an increase in the columnar organization. There was no significant association between cortical cataract (C) and this feature. The number of black holes was weakly and inversely associated (P = 0.016) with the LOCS ΙΙΙ C grade. However, the size of the black holes showed no association with any of the LOCS III classes. 

Table 2 shows the number of right and left eyes (OD and OS) in each range of mean cell density. There is an inverse association of high statistical significance between the average cell count and age (P = 0.014); the annual decrease was 9 cells/mm2. This statistically significant association between average cell density and age persisted even when the generalized linear model with intraclass correlation regression models were controlled for sex, size or number of black holes, or LOCS ΙΙΙ grade of cataract (see Table 3 ). The grade of the feature puffy clouds was highly significantly and inversely associated with the LEC count (P = 0.0006). In the regression model controlled for age, sex, and LOCS III grade, the regression coefficient indicated that there was a reduction of 175 cells/mm² with each step increase in the feature puffy clouds. It is not known whether this feature bears any association with increased cell size and swelling with age, noted in other studies. ^{4 5} This feature showed no association with LOCS III classes C, P, NO, or NC. 

A significant inverse association existed between the cell count and both the size (P = 0.0013) and number (P = 0.0000) of black holes (Table 3) . This association persisted in generalized linear model with intraclass correlation models controlled for age, sex, and LOCS III classification (Table 4) . The grade of the feature linear furrow showed no association with either the LOCS III class or LEC count. We do not know the significance of this feature at this time. The relation of average cell count to the LOCS III grade of ARC was not statistically significant. 

The histology of the central zone of the lens epithelium from the representative continuous curvilinear capsulorrhexis specimens is shown in Figures 3 and 4 . The central zone of epithelium, irrespective of cataract type, was not a uniform cuboidal monolayer. Rather, it was found to consist of segments of variably high cuboidal or low columnar, cuboidal, low cuboidal, and squamous epithelial cells. The cells within these segments also exhibited different levels of nuclear and cytoplasmic basophilia. Low columnar and cuboidal cells were generally lightly stained, whereas the low cuboidal and squamous cells were consistently more heavily stained. In cross-sections of the epithelium the cell heights varied, and in serial sections there appeared to be alternating circumscribed areas of high and low cells which we believe correspond to the puffy clouds, variable columnar organization, and linear furrows observed and quantified in specular images (see Table 3 and Fig. 2 ). In addition, small regions of the central zone were observed that were either denuded of epithelial cells (Fig. 3D) or populated by extremely flattened squamous cells (Fig. 4D) . These regions were observed as a function of age, strongly suggesting that these are the black holes detected by specular microscopy. 

In thin sections, the irregular heights and shapes of central zone epithelial cells were confirmed, and ultrastructural features indicative of variable age were revealed. The squamous and low cuboidal cells (Figs. 5A 5B ) were generally vesiculated, showed no defined cytoskeletal structures or latticework, and had infrequent mitochondria with few cristae and nuclei with condensed chromatin. In addition, neighboring cells within these regions had variable degrees of cytoplasmic basophilia that enhanced the appearance of their highly irregular lateral and apical membranes. All these feature are typical of old or senescent cells. ²⁶  

In contrast, cuboidal and low columnar cells were more uniformly basophilic and had more mitochondria, with full complements of cristae. These cells frequently had an organized apical cytoskeletal network ^{27 28 29} and generally had nuclei with less condensed chromatin. All these features are typical of nonsenescent cells. However, at the ultrastructural level, it was seen that the shape of these cells was often neither truly cuboidal nor columnar. Rather, they were groups of variably upright to inverted pyramidal cells with little if any interlocking lateral membranes (Fig. 5C) . In addition, there were regions of pseudostratified and stratified epithelium. The pseudostratified epithelial cells often appeared to be stratified (Figs. 6A 6B ). However, in serial sections, it was revealed that all cells had at least a portion of their basal membranes in contact with the capsule. Clearly, cell counts from these regions, in either specular images or wholemounts of continuous curvilinear capsulorrhexis specimens, could easily underestimate total cell counts. There were also examples of cuboidal cells with markedly tortuous lateral interdigitating membranes (Fig. 7A ) and also examples of cuboidal cells with rounded lateral membranes extending from the apical membrane (Fig. 7B) . These cells were frequently seen adjacent to regions ostensibly denuded of epithelial cells. 

Regardless of the specific morphologic change(s) in the epithelium, which created the black holes, their presence indicates an abrupt change in the reflectivity of the epithelial surface in that area. The gaps between and the attenuation of the LEC evident in electron microscopy are possible sources of this altered reflectivity. The puffy clouds correspond well to areas of pseudostratified and stratified epithelium. 

Conclusions

The enhanced specular images obtained were suitable for measuring the LEC density and for identifying and grading four major features of the image. Our results showed conclusively that there was a statistically significant decline in cell density with age. We believe that our study is the first to show this decline with regression models controlled for gender, cataract type, and severity. Earlier studies cited in our introduction, showed a decline in LEC density with age or cataract, but did not control the measure for all other confounding factors. It is not surprising that the LEC density decreased with the size and number of black holes. As the area in the specular image occupied by the black holes increased, a lower cell density would be expected. However, the decline in cell density with age persisted in models controlled for both the size and the number of black holes. This finding suggested that the decline in LEC density with age was not artifactual. Assuming that this rate of decline in LEC density was linear, there would be a loss of 675 cells/mm² in a 75-year life span, a loss of 14% of the total number of LECs. Our estimate of the annual decrement in the cell density corresponded almost exactly to that observed by Guggenmoos–Holzmann et al., ⁷ based on their study of flat preparations of central lens epithelium obtained from donor eyes. The lack of longitudinal data on LEC density makes the estimate of lifetime loss purely speculative, but this speculation suggests that this loss may not be inconsequential. We found also that the decline in the cell density was significantly and inversely associated with the severity of the feature puffy clouds. That there was no statistically significant association between puffy clouds and the LOCS ΙΙΙ grade made it difficult to infer a relationship between decreased cell density and ARC. Clearly, there was something more than a decline in cell number that led to ARC formation. 

We also studied the association of the other three major features of the specular image with the severity of cataract. Only the feature columnar organization was associated inversely with increasing severity of P and NO cataract. With increasing severity of P or NO, there was a reduction in the columnar organization of the LECs. This reduction is consistent with the increased pleomorphism of LECs noted in cataractous lenses in studies in vitro. 

We have at present no way to infer the metabolic activity or altered cell function in aging LECs in vivo, and we therefore cannot relate these anatomic data to existing physiological data on LECs. 

We are just beginning to understand the cellular anatomy responsible for the changes noted in the specular images of the LECs. Correlative light and electron microscopic analysis revealed that the puffy clouds in specular images were circumscribed areas of either pseudostratified or stratified epithelium. In the same manner, black holes seen in specular images were either areas of extremely flattened and aged squamous epithelium or infrequent small areas devoid of epithelial cells. The cells adjacent to sites devoid of epithelial cells always had rounded lateral membranes. This feature is consistent with these cells spreading to cover a site previously occupied by a cell eliminated by apoptosis. The striking specular image feature black hole, confirmed earlier data from Fagerholm and Philipson ⁹ who showed areas of missing cells in lenses with pronounced cortical swelling, vacuolization, and subcapsular opacification. However, there is controversy regarding the nature of cell death in the lens epithelium. Both apoptosis and necrosis have been cited as the cause of LE cell death in recent publications. ^{30 31}  

The role of the noncontact corneal specular microscope in the clinical assessment of cataract is yet to be defined. Certainly, it provides an image that when enhanced with inexpensive, commercially available image-processing software, clearly demonstrates the features of the human lens epithelial layer. It is now possible to measure LEC density noninvasively and to grade severity of features in specular images. The noncontact corneal specular microscope can be incorporated into epidemiologic studies of cataract and the risk factor profiles for specific morphologic changes in the lens epithelium identified. 

Our inability to find striking associations between LEC density, specular image severity, and the LOCS ΙΙΙ grade of cataract, suggests that more study of these issues is needed. Shifting the focus to younger populations with less overt cataract to groups of individuals with well-defined risk factor profiles for cataract and to assays of lens function (rather than assays of lens morphology) may help establish the role of noninvasive specular imaging of lens in the arena of cataract research. 

 Supported by an unrestricted grant from Konan, Inc., Hyogo, Japan; a grant from the Massachusetts Lions Eye Research Fund; and Grants RO1 EY-12015 and RO1 EY-06642 from the National Eye Institute.

 Submitted for publication April 16, 1999; revised September 3, 1999; accepted September 23, 1999.

 Commercial relationships policy: C2.

 

Figure 1.

 

Specular microscope for the lens (SML-2; Konan, Hyogo, Japan).

Table 1.

 

Number of Eyes by LOCS III Grade

LOCS III Class	Grade									Total
		0.1–1.0	1.1–2.0	2.1–3.0	3.1–4.0	>4.0
C (Cortical)	150	38	40	26	0	254
P (Posterior subcapsular)	214	26	8	6	0	254
NO (nuclear opalescence)	42	76	96	26	12	252
NC (nuclear color)	48	86	70	30	18	252

Figure 2.

 

Enhanced specular images of human lens epithelium showing the features that were graded. (A) The linear furrows are black, spindle- or furrow-shaped areas marked by the dot in the horizontal i of the top two images. (B) The term columnar organization refers to the vertical pattern of rows marked by the dot on the horizontal i. (C) The term puffy clouds refers to the whitish cloudlike areas marked by the dot on the horizontal i. (D) The size of the black holes varies in this set of images, and in (E) the number of black holes varies. Paired images show lowest and highest grades of each feature. Bar, 50 μm.

Table 2.

 

Number of Eyes in Each Range of Average Lens Epithelial Cell Density

Eye	n	≤3600*	>3600 to ≤4000	>4000 to ≤4500	>4500 to ≤5000	>5000 to ≤5500	>5500
Right	139	12	22	44	49	10	2
Left	120	29	24	35	20	8	4

*  Cells per square millimeter.

Table 3.

 

Association between LOCS III Grade, LEC Count, and Severity of Specular Image Features

LOCS III Class*	Feature of Specular Image of Lens Epithelium
		Furrows	Columnar	Clouds	Hole Size	Number of Holes
C
P	NS	NS	NS	NS	0.016
β, †					−0.15
β, † /SE					−2.42
P
P	NS	0.001	NS	NS	NS
β, †		−0.13
β, † /SE		−3.24
NO
P	NS	0.003	NS	NS	NS
β, †		−0.08
β, † /SE		−3.20
NC
P	NS	0.0003	NS	NS	NS
β, †		0.12
β, † /SE		3.65
LEC count (average)
P	NS	NS	0.0006	0.0013	0.00000
β, †			−175	−223	−283
β, † /SE			−3.43	−3.23	−4.85

 All models controlled for age and sex. NS, not significant (P > 0.05).

*  Dependant variable.

†  Regression coefficient.

Table 4.

 

Regression Modeling Showing Association between Age and the Average LEC Count

Independent Variable of Interest	Independent Variables	β* P	Average Cell Count
Age	Size of black holes, sex	β	−8.27
		p value	0.014
Age	Number of black holes, sex	β	−7.72
		p value	0.017
Age	LOCS III C, sex	β	−9.06
		p value	0.007
Age	LOCS III P, sex	β	−8.08
		p value	0.014
Age	LOCS III NO, sex	β	−8.57
		p value	0.013
Age	LOCS III NC, sex	β	−8.61
		p value	0.014

*  Regression coefficient.

Figure 3.

 

Light microscopy of central capsulorrhexis specimen showing nonuniformity in the cell monolayer (A, B, and C) and areas of cell dropout (D).

Figure 4.

 

Light microscopy of central capsulorrhexis specimen showing segments of lightly stained low columnar and cuboidal cells and heavily stained low cuboidal and squamous cells (A, B, and C). Some segments consisted of extremely flattened squamous cells (D). Scale bar (A and B) = 20μ m; (C and D) = 10 μm.

Figure 5.

 

Electron micrographs showing differences in ultrastructural features between different cell types. The squamous and low cuboidal cells have increased vesiculation and do not have a well-defined cytoskeleton (A, B). The cuboidal and low columnar cells show organized cytoskeleton and less condensed chromatin in their nuclei (C).

Figure 6.

 

Segments of the central capsular zone showing similarity between pseudostratified (A) and stratified epithelium (B). Serial sections revealed that all cells in (A) had a portion of their basal membranes in contact with the capsule.

Figure 7.

 

Electron micrograph showing cuboidal cells with markedly tortuous interdigitating lateral membranes (A) and cells with rounded lateral membranes extending from the apical membrane (B).

The authors thank Bernard Rosner (Channing Laboratory, Harvard Medical School) for assistance with statistical analysis.