Corneal Opacity in Lumican-Null Mice: Defects in Collagen Fibril Structure and Packing in the Posterior Stroma

Abstract

purpose. Gene targeted lumican-null mutants (lum ^tm1sc /lum ^tm1sc ) have cloudy corneas with abnormally thick collagen fibrils. The purpose of the present study was to analyze the loss of transparency quantitatively and to define the associated corneal collagen fibril and stromal defects.

methods. Backscattering of light, a function of corneal haze and opacification, was determined regionally using in vivo confocal microscopy in lumican-deficient and wild-type control mice. Fibril organization and structure were analyzed using transmission electron microscopy. Biochemical approaches were used to quantify glycosaminoglycan contents. Lumican distribution in the cornea was elucidated immunohistochemically.

results. Compared with control stromas, lumican-deficient stromas displayed a threefold increase in backscattered light with maximal increase confined to the posterior stroma. Confocal microscopy through-focusing (CMTF) measurement profiles also indicated a 40% reduction in stromal thickness in the lumican-null mice. Transmission electron microscopy indicated significant collagen fibril abnormalities in the posterior stroma, with the anterior stroma remaining relatively unremarkable. The lumican-deficient posterior stroma displayed a pronounced increase in fibril diameter, large fibril aggregates, altered fibril packing, and poor lamellar organization. Immunostaining of wild-type corneas demonstrated high concentrations of lumican in the posterior stroma. Biochemical assessment of keratan sulfate (KS) content of whole eyes revealed a 25% reduction in KS content in the lumican-deficient mice.

conclusions. The structural defects and maximum backscattering of light clearly localized to the posterior stroma of lumican-deficient mice. In normal mice, an enrichment of lumican was observed in the posterior stroma compared with that in the anterior stroma. Taken together, these observations indicate a key role for lumican in the posterior stroma in maintaining normal fibril architecture, most likely by regulating fibril assembly and maintaining optimal KS content required for transparency.

It has long been recognized that collagen architecture of the corneal stroma is crucially important in the ultimate transparency of the cornea. ¹ Collagen fibrils in the stroma are maintained in the range of 20 to 40 nm and organized into a highly ordered, latticelike configuration. The highly ordered architecture of the corneal stroma is affected by multiple factors. Recent studies of several types of hereditary corneal dystrophies elucidated abnormal collagen fibril architecture of the corneal stroma. For example, corneal opacification is a clinical feature of Scheie’s syndrome or mucopolysaccharidosis (MPS) type I, a lysosomal storage disorder with an iduronidase A deficiency. ² In addition to featuring granular deposits, transmission electron microscopy of MPS I–affected corneas revealed the presence of thicker collagen fibrils and localized disorganization of the matrix. ^{3 4} Macular corneal dystrophy, with deficiencies in keratan sulfate (KS) biosynthesis, also causes clouding of the cornea and similar disruptions in stromal fibril structure and organization. ^{5 6 7 8} In both cases, altered proteoglycan synthesis and composition are to be expected. Recently, a mouse model for corneal dystrophy was developed by targeted disruption of the lumican gene (lum ^tm1sc /lum ^tm1sc ). ⁹ The mutant mice had cloudy corneas and stromal collagen fibrils with increased diameter and altered structure. 

Lumican is a member of the leucine-rich proteoglycan (LRP) family. ¹⁰ It is a major keratan sulfate proteoglycan of the corneal stroma as well as other collagenous extracellular matrices (skin, cardiac valves, cartilage, and bone). ¹¹ Other LRP members include decorin, fibromodulin, biglycan, keratocan, osteoglycin, and epiphycan. ¹² Decorin, a chondroitin sulfate (CS) proteoglycan widely expressed during mouse embryonic development, is also a major component of the corneal stroma. ¹³ Previous studies have shown that the core proteins of lumican, decorin, and other LRPs from tendons can delay spontaneous collagen fibril formation and inhibit the lateral growth of fibrils in fibrillogenesis assays in vitro. ^{14 15 16} Also, the abnormal lateral growth of isolated corneal fibrils stripped of their surface-associated macromolecules is prevented by the corneal proteoglycans. ¹⁷ Recent gene-targeting studies of LRPs suggest a similar role for these proteoglycans in vivo. Thus, absence of lumican in our lum ^tm1sc /lum ^tm1sc mouse model of corneal dystrophy affected collagen architecture of the cornea and skin with consequent corneal opacity and reduced dermal biomechanical tensile strength. In addition to lumican, gene-targeted null mutations in decorin and fibromodulin also led to abnormal collagen fibril architecture in skin and tendons. ^{18 19} However, to date only the lumican-deficient mice have demonstrated a corneal phenotype. 

The purpose of the present study was to assess corneal opacification in the lum ^tm1sc /lum ^tm1sc mice and define its source in the corneal stroma by in vivo confocal microscopy. Parallel analyses of collagen fibril structure, fibril packing, and organization in the lumican-deficient and wild-type control mice and lumican expression in the mature normal cornea indicate that lumican serves a key role in the establishment and maintenance of corneal transparency. 

Methods

Animal Husbandry

All experiments were performed in compliance with the ARVO Statement for Use of Animals in Ophthalmic and Vision Research. CD-1 outbred wild-type and lum ^tm1sc /lum ^tm1sc gene-targeted mice in the same genetic background were housed in a specific pathogen-free mouse housing facility at Case Western Reserve University. For the in vivo confocal microscopic through-focusing (CMTF) analyses, live mice were shipped to the Southwestern Medical Center at Dallas. 

In Vivo Confocal Microscopy

In vivo examinations were performed using a scanning confocal microscope with a z-axis resolution of 9 μm (Tandem Scanning, Reston, VA). Before examination, mice were anesthetized with an intramuscular injection of ketamine HCl (50 mg/ml), xylazine HCl (20mg/ml), and acepromazine (10 mg/ml) in sterile water. A drop of 2.5% hydroxypropyl methylcellulose was then applied to the objective tip as an immersion fluid, because the objective was brought into contact with the mouse eye. With the gain, kilovolts, and black level on the video camera kept in automatic mode, initially high-quality images were obtained from all corneal sublayers. These were then switched to manual settings that were kept constant throughout the study, to allow for direct comparison of scans obtained at different time points. Digital image acquisition and CMTF have been described earlier. ^{20 21}  

Conventional Transmission Electron Microscopy

Corneas of lum ^tm1sc /lum ^tm1sc and wild-type controls of similar age (lum ⁺ /lum ⁺) were fixed in 4% paraformaldehyde, 2.5% glutaraldehyde, and 0.1 M sodium cacodylate, (pH 7.4) with 8.0 mM CaCl₂ for 2 hours on ice. The corneas were dissected and postfixed with 1% osmium tetroxide for 1 hour. After dehydration in a graded ethanol series followed by propylene oxide, the corneas were infiltrated and embedded in a mixture of Polybed 812, nadic methyl anhydride, dodecenylsuccinic anhydride, and 2,4,6-tris(dimethylaminomethyl)phenol (DMP-30; Polysciences, Warrington, PA). Thick sections (1 μm) were cut and stained with methylene blue-azur blue for examination and selection of specific regions for further analysis. Thin sections (100 nm) were prepared using an ultramicrotome (UCT; Reichert Jung, Vienna, Austria) and a diamond knife and stained with 2% aqueous uranyl acetate followed by 1% phosphotungstic acid (pH 3.2). The sections were examined and photographed at 80 kV using a transmission electron microscope (model 7000; Hitachi, Tokyo, Japan). The microscope was calibrated using a line grating. 

Fibril Diameter Measurements

The corneal stroma was divided into three regions for analysis: anterior, mid and posterior stroma. The anterior stroma was defined as the 10 μm subjacent to the epithelium, and the posterior stroma was the 10 μm adjacent to Descemet’s layer. When specifically stated, regions of the posterior stroma were chosen for analysis based on the presence of structurally abnormal fibrils. In all other experiments calibrated micrographs from each region were chosen randomly in a masked manner. Micrographs of appropriate regions were taken at× 48,700. The micrographs were digitized, and diameters were measured by computer (RM Biometrics-Bioquant Image Analysis System; Nashville, TN). 

Immunofluorescence Microscopy

Corneas from 3-month-old and more than 7-month-old wild-type and lumican-deficient mice were fixed with 4% paraformaldehyde in phosphate-buffered saline (PBS; pH 7.3) for 30 minutes on ice. The tissues were cryoprotected with 2 M sucrose-PBS, and frozen in optimal cutting temperature compound (OCT; Tissue Tek, Miles Laboratories, Naperville, IL). Sections (6 μm) were cut and mounted on poly-l-lysine–coated slides. Sections were treated with sodium borohydride (50 mg/100 ml PBS) for 15 minutes at room temperature, and nonspecific binding sites were blocked by incubation in 5% bovine serum albumin (BSA) in PBS overnight at 4°C. Sections were then incubated with a polyclonal anti-mouse lumican antiserum (1:250) obtained from Ake Oldberg (University of Lund, Sweden). The lumican antiserum was generated against a bacterially produced lumican fusion protein and characterized as lumican specific. ¹⁹ The sections were then incubated with a secondary goat anti-rabbit dichlorotriazynyl amino fluorescein–conjugated antibody (1:150; Jackson ImmunoResearch, West Grove, PA). Negative control samples were incubated identically, except the primary antibody was excluded. To visualize nuclei, the slides were mounted in glycerol solution with 1 mg/ml Hoechst stain. Images were captured using a digital camera (Optronics, Goletta, CA), with set integration times and identical conditions to facilitate comparisons between samples. However, for optimal presentation of the data, images presented had exposure times as indicated in the figure legend. 

KS and CS Content

Glycosaminoglycan (GAG) content per whole eye was determined by modifying a previously published method that uses the ability of proteoglycans to bind the dimethylmethylene blue (DMB) dye. ²² Whole eyes were extracted in 0.05 M sodium acetate containing 4 M guanidium-HCl (GuHCl). The soluble supernatant containing proteoglycans was dialyzed against 0.1 M Tris (pH 7.4) containing phenylmethylsulfonyl fluoride (500 mM), N-ethylmaleimide (100 mM), and EDTA (100 mM) and incubated with DMB reagent prepared as described. The change in absorbance was determined in an enzyme-linked immunosorbent assay reader (model II ELISA reader; Dynatech, Alexandria, VA) as a ratio of 550:610 nm. To estimate the amount of KS and CS in the total GAG content, the extracts were digested with either keratanase or chondroitinase ABC (Seikagaku, America, Rockville, MD) to specifically remove KS or CS, and DMB-reactive GAGs were estimated as described. KS or CS in micrograms per eye was calculated as follows. Commercially available KS and CS (Seikagaku America) at various concentrations (0–10 μg/well) was mixed with DMB, and absorbance at 550:610 nm was determined to generate a standard curve to estimate proteoglycan-DMB binding. Absorbance at 550:610 nm of whole-eye extracts reacted with DMB, before and after enzymatic removal of GAGs, was compared with the standard curve to estimate micrograms of proteoglycan per eye. 

Results

In Vivo Confocal Microscopy

Corneas were examined by in vivo CMTF and CMTF profiles typical of corneas from lum ⁺ /lum ⁺ and lum ^tm1sc /lum ^tm1sc mice were compared ( Fig. 1 ). Unlike the corneas of wild-type mice, the entire stroma in the lumican-deficient mice appeared bright due to increased backscattered light. In addition, there was a distinct, highly reflective zone in the posterior one third of the stroma adjacent to the endothelium. This reflective zone was clearly absent in the wild-type corneas. 

Stromal backscattering was calculated as the difference between stromal pixel intensity and the background pixel intensity (the plateau in intensity level after the endothelial peak; Table 1 ). Stromal backscattering in the lumican-deficient mice was significantly greater than in age-matched control animals. The lumican-null mice, on average, displayed a threefold increase in normalized backscattering of light over control mice in the two age groups examined (P < 0.05). 

The profiles of the scanned images were also used to determine thickness of the corneal sublayers (Table 1) . In the 4- to 5-month-old mice stromal thickness in lumican-null mutants and wild types was 44.1 ± 7.3 and 74.3 ± 11 μm, respectively. This reflected a significant 40% reduction in stromal thickness in the lumican-null mutants (P < 0.0005). The 1- to 2-month-old lumican knockout mice displayed similar reduction in stromal thickness (P = 0.0002). Therefore, when we normalized backscattered light by dividing by stromal thickness, the difference between lumican-deficient and normal control corneas was even more significant (Table 1) . Epithelial thickness in the mutant mice was comparable to that in wild-type mice in the 1- to 2-month age group. However in the 4- to 5-month group there was a small (12%) but significant increase in epithelial thickness in the mutant mice. 

Abnormal Collagen Fibril Architecture Localizes to Posterior Stroma in lum^tm1sc/lum^tm1sc Mice

Corneas from adult mice (>7 months old) were analyzed by transmission electron microscopy. The stroma was divided into three regions for analysis: anterior, mid, and posterior. In lumican-deficient and wild-type corneas, the fibrils in the anterior stroma had regular packing and cylindrical collagen fibrils. In contrast, in the posterior stroma of the lumican-deficient mice, fibril diameters were larger. In addition, there were fibrils laterally associated with one another and fibrils with irregular contours (Fig. 2) . The midstromal region was unremarkable in fibril structure, except for the rare appearance of abnormal laterally fused fibrils typical of the posterior stroma. The midstroma closer to the posterior stroma had a higher incidence of these abnormal fibrils (data not shown). Packing of the collagen fibrils in the lumican-null mutants was fairly unperturbed in the anterior stroma but was affected severely in the posterior stroma, with frequent fibril-poor spaces or lakes. ²³  

Abnormal Morphology of lum^tm1sc/lum^tm1sc Collagen Fibrils

The structure of fibrils in the posterior stroma of the lumican-deficient corneas was analyzed in greater detail (Fig. 3) . The posterior stroma (10 μm adjacent to Descemet’s layer) contained a distinct population of fibrils with very large diameters, and a subset of these had irregular contours in addition to laterally associated fibrils (Fig. 3A , arrows; 3B, 3C, 3D, higher magnification showing irregular fibril contours). A longitudinal view of a large-diameter fibril is shown in Figure 3E . These structural alterations are indicative of a defect in the regulation of fibril growth—i.e., an abnormal lateral growth not normally seen in the corneal stroma. 

Fibril diameter distributions were determined for the anterior and posterior stroma of lumican-deficient and age-matched wild-type control corneas (Fig. 4) . Fibril morphology in the anterior stromas from both corneas was virtually identical (Fig. 4A) . Although we noted a consistent and reproducible increase in mean fibril diameter in lumican-deficient versus wild-type control, the difference was not statistically significant. In contrast, in the posterior region of the stroma there was a significant increase (P = 0.005) in mean fibril diameter: 35.4 nm versus 37.6 nm for wild-type and lumican-deficient corneas, respectively. In addition, the distribution of fibril diameters showed a clear shift toward larger diameters in the mutant corneas. This shift in distribution included a distinct population of larger fibrils not seen in the normal control animals (Fig. 4B ; arrows electron micrograph). These measurements, however, are a characterization of fibril diameters in the posterior stroma, with fields being chosen randomly without specifically including areas rich in structurally abnormal fibrils. 

Chakravarti et al. ⁹ in an earlier study on lumican-deficient mice reported mean fibril diameter (47.3 nm) and diameter distributions (range, 20–235 nm) that were much higher than those reported in the present study in which randomly selected fields from the posterior stroma were used. This apparent discrepancy stems from the fact that the posterior stroma is quite heterogeneous in numbers and distribution of abnormal fibrils, and the previous study compared these localized, large, fibril-rich areas to the wild-type posterior stroma. When fibrils were analyzed from preselected abnormal fibril-rich areas, similar values for mean fibril diameter (51.2 ± 23 nm) and range (19–136 nm) were obtained. 

Abnormal Fibril Packing and Lamellar Organization in Posterior Stromas of lum^tm1sc/lum^tm1sc Mice

In addition to the defects in fibril structure, the posterior stroma of the lumican-deficient corneas harbored alterations in interfibrillar spacing associated with abnormal fibril packing and a dramatic disruption in lamellar organization of fibrils. These features become apparent when corneal sections were viewed under low magnification from the anterior to the posterior stroma (Fig. 5) . Relative to lamellae in wild-type controls (bold arrows in Fig. 5A ), lamellae in the posterior stroma of the mutant corneas (Fig. 5B) were poorly organized and difficult to differentiate from adjacent lamellae. In particular, fibrils lost their regular packing adjacent to the Descemet’s layer (asterisks, Figs. 5B 5C ). In the anterior stroma, lamellar disorganization was less apparent, and approximately 10 to 15μ m away from Descemet’s layer approaching the anterior stroma, lamellar architecture was relatively normal. 

Lumican Distribution in the Cornea

Structural and functional defects clearly localized to the posterior stroma in lumican-deficient mice. Therefore, we assessed immunohistochemically the relative distribution of lumican in the normal cornea. In young adults (3 months of age), lumican immunoreactivity present throughout the stroma was enhanced markedly in the posterior third of the stroma (Figs. 6A 6B 6C ). In the mature adult (>7 months of age), although this posterior localization was maintained, there was an increase in reactivity throughout the stroma with a consequent reduction in the sharp gradient seen in the 3-month-old animals (Fig. 6D) . The posterior enrichment in lumican reactivity coincided with the regions demonstrating structural and functional defects described earlier. In the lumican-deficient cornea there was no lumican reactivity as expected (Fig. 6E) . The epithelium of both wild-type and lumican-null mice showed low levels of nonspecific staining (Figs. 6B 6E) . Of interest, compared with wild-type adults, staining for keratocan and decorin was unchanged in the mutant corneas (data not shown). Thus, there did not seem to be a compensatory overexpression of these LRPs in the lumican-null corneal stroma. 

KS and CS Content of Eyes

KS has been historically viewed as essential for corneal transparency. Absence of lumican, a major KS proteoglycan (KSPG) of the cornea, may lower the KS pool, which can be a factor contributing to corneal opacification in knockout mice. To test this possibility we determined KS and CS content of whole eyes in wild-type and lumican-deficient mice. Mean KS content per eye for wild-type mice was 1.34 ± 0.09 μg and for lum ^tm1sc /lum ^tm1sc was 1.03 ± 0.05 μ g. The latter reflected a significant decrease of 25% in KS content (P = 0.01). It was not surprising that mean CS content in the wild-type mice (1.064 μg per eye) and the lumican-deficient mice (1.051 μg per eye) was not different (P = 0.93). 

Discussion

The cornea is the outermost transparent protective barrier that provides 70% of the refractive power of the eye. The corneal stroma elegantly exemplifies the basic biologic principle that structure defines function. It was recognized very early that for the cornea to be transparent, it must have a highly ordered latticelike collagen fibril architecture with uniformly thin diameter and regular interfibrillar spacing. ¹  

The lumican-deficient mouse model links abnormal collagen architecture with loss of corneal transparency, providing the first genetic mouse model for corneal opacity. ⁹ Mutant mice were assessed qualitatively for clouding by slit lamp biomicroscopic examination of the corneas. Loss of corneal transparency could be detected as early as 5 weeks. In the present study, in vivo CMTF was used to obtain a quantitative assessment of corneal opacity. Backscattered light can be used as an indication of the extent of corneal opacity. Regional differences in the quantity of backscattered light can provide additional information on the status of corneal transparency across the entire thickness of the cornea. This technique has been used on a rabbit model for wound repair after photorefractive keratectomy. ²⁴ Three-dimensional reconstruction of the CMTF scans of the mouse cornea in our study showed a distinct highly reflective zone in the posterior stroma. Profiles of the scanned images confirmed a more than threefold increase in backscattered light in the lumican-deficient mice. 

The CMTF data provided the expected indication of loss of transparency long before the loss was detected by slit lamp examination. ⁹ However, it was not expected that the intensity of backscattered light in the younger mice would be as high as it was in the older mutants. Given the extreme sensitivity of CMTF scanning, the backscattering in the lumican-deficient mice examined in the current study may already be at saturation levels. Earlier examinations of the same animal over time, from postnatal to adult stages, will provide further insight on the progressive nature of corneal opacity. 

The analysis of corneal opacification clearly indicated that the defect resides in the posterior stroma. Therefore, we carefully examined the collagen architecture in the mutant mice, from the anterior stroma containing newly synthesized collagen fibrils to the posterior stroma containing mature collagen fibrils. The architectural alterations coincided with the functional defect, in that the anterior stroma was normal in fibril diameter and packing, but in the posterior stroma, fibril diameter had increased by 6%. The posterior stroma also contained a population of large-diameter fibrils, many with irregular contours indicative of abnormal lateral growth of fibrils. ²⁵  

A recent study also reported irregular large fibrils in the posterior stroma in lumican-deficient mice. ²⁶ Abnormal lateral growth requires lateral associations of collagen fibrils that would disrupt fibril packing and organization. Both packing and general lamellar organization of the fibrils were dramatically disrupted in the posterior stroma of the lumican-deficient mice. These structural defects in the lumican-null mice agree with our immunohistologic results that demonstrate a much stronger presence of lumican in the posterior stroma of wild-type normal corneas. An almost total restriction of the fibril- and matrix-disruptions to the posterior stroma of the mutant mice implies a unique role for lumican in this region. Lumican-collagen interactions in the posterior stroma may be of particular significance in limiting interfibrillar associations between mature fibrils. Perhaps other KS-LRPs usurp this role in other regions of the stroma. Certainly, the heterogeneity of the defect also suggests other regulators of stroma matrix assembly that remain to be elucidated. The increase in fibril diameter and thinning of cornea that we have reported in lum ^tm1sc /lum ^tm1sc mice occurs in several types of mucopolysaccharidosis ^{3 25 27} and keratoconus, ²⁸ respectively. However, the massive matrix disorganization of the posterior stroma in lumican-deficient mice was not seen in the human dystrophies. 

The structural and functional differences in the anterior versus the posterior stroma of the lumican-null mutants should be reviewed in the context of the differences between the anterior and posterior stroma noted previously. ^{29 30 31 32} In agreement with our finding that the posterior stroma is rich in lumican is the observation that this region of the stroma also has a higher concentration of KS than the anterior stroma. Our results for KS estimation in whole eyes detected a lowering of total ocular KS levels in the lumican-null mutants. We speculate a similar reduction of KS from the posterior stroma in the mutants may contribute to the anomalies seen in these mutants. It was pointed out earlier that low oxygen content in the posterior stroma allowed efficient synthesis of KSPG and not CS proteoglycan (CSPG), which is oxygen-dependent. ³³ Yet, even the very thin mouse cornea, less than 20% the thickness of rabbit and bovine corneas, is rich in KSPG in the posterior stroma, indicating an intrinsic need for KSPG in this region. Studies on the bovine cornea show increased hydration in the posterior stroma, and KSPG is reported to have higher water retention capability. Thus, it is likely that lumican in the posterior stroma is required for a KS-rich environment to maintain a highly hydrated state. 

In addition to lumican, keratocan and mimecan are two major KSPGs of the cornea. ^{34 35} Yet, these are not sufficient to complement the contributions of lumican to maintenance of stromal architecture. It is possible that their contribution to the KS pool in the posterior stroma is not high enough to compensate for a reduction in KS associated with the absence of lumican. Ocular phenotypes of the keratocan- and mimican-deficient mice will provide further clues about their role in the cornea. Keratocan, specifically expressed throughout the cornea during development and in adult mice, may adequately maintain anterior stromal architecture as well as stromal physiological properties. ³⁶ Mice deficient in keratocan may have corneal developmental anomalies, anterior stromal collagen defects, and changes in corneal physiology. The heterogeneity in the posterior stroma of the lumican-deficient corneas implies that keratocan and/or mimican also have a secondary role in this region. 

The molecular mechanism underlying the considerable thinning of the stroma in the lumican-null mice remains to be elucidated. It may be due to a smaller pool of stromal keratocytes in the stroma, implying that lumican may play a role in the early stages of cell migration into the developing stroma. Alternatively, absence of lumican in the null mice may have an adverse effect on the biosynthesis of other stromal components. In the 4- to 5-month age group we also noted a small increase in the epithelial thickness in the lumican-null mice. Recently, the corneal epithelium has been shown to express lumican after wounding, indicating a biologic role for lumican in the epithelium as well. ²⁶ Thus, it is possible that the observed thickening of the epithelium in our older lumican-null mice may be a compensatory wound-healing response to lumican-deficiency and the thinner structurally impaired stroma. 

The present study showed lumican to be present in high quantities in the posterior stroma. The collagen structural and organizational anomalies observed in its absence in the posterior stroma of lumican-null mutants establishes a key role for lumican in maintaining the unique properties of the posterior stroma. 

 Supported by National Institutes of Health Grants EY11654 (SC), EY05129 (DEB), Research to Prevent Blindness (JHL), EY 11373 (JHL), and EY07348 (JVJ).

 Submitted for publication February 23, 2000; revised May 18, 2000; accepted June 16, 2000.

 Commercial relationships policy: N.

 

Figure 1.

 

In vivo CMTF three-dimensional image and profile of lum ⁺/lum ⁺ and lum^tm1sc/lum^tm1sc. Three-dimensional CMTF images of typical corneas from 5-month-old wild-type (A) and lumican-null mutants (B) show the stroma (b and b′) spanned by an epithelium (a and a′) and the endothelium (c and c′). The backscatter in the mutant corneal stroma increased maximally near the endothelium. A traced profile of the scanned images (C) shows a marked increase in the intensity of backscattered light in the mutant stroma (b′) with a sharp increase approaching saturation at the endothelium (c′).

Table 1.

 

Epithelial, Stromal, and Corneal Thicknesses and Light Scattering in the Lumican-Null (lum ^tm1sc /lum ^tm1sc ) Mouse

Mouse Strain	Age (mo)	Thickness (μm)							Backscattering
				Eyes (n)	Epithelium	Stroma	Cornea	Eyes (n)	Intensity	Normalized
lum +/lum +	1–2	4	39.0 ± 4.5	73.0 ± 3.7	112.1 ± 4.8	4	2787 ± 626	38 ± 9
	4–5	10	41.0 ± 2.3	74.3 ± 11.2	115.3 ± 10.3	10	2344 ± 148	33 ± 6
lum tm1sc /lum tm1sc	1–2	6	36.6 ± 4.3	39.3 ± 7.3	75.8 ± 7.8	6	3933 ± 691	102 ± 24
	P*		NS	<0.0005	<0.0005		<0.05	<0.005
	4–5	8	46.6 ± 3.8	44.1 ± 7.3	90.7 ± 7.0	4	4441 ± 1404	95 ± 15
	P		<0.005	<0.0005	<0.0005		<0.05	<0.05

 Backscattering was normalized to stromal thickness. NS, not significant.

*  Compared with control eyes; P based on Student’s t-test or Mann–Whitney test.

Figure 2.

 

Fibril defects are localized to the posterior corneal stroma in corneas from 7.5-month-old lum ^tm1sc /lum ^tm1sc mice. Transmission electron micrographs comparing collagen fibril structure in the anterior (A, B) and posterior (C, D) stroma of lumican-deficient (B, D) and wild-type (A, C) corneas. Fibril structure and packing are comparable in the anterior stroma of wild-type (A) and null mice (B). In contrast, the fibrils in the posterior stroma of lumican-deficient mice (D) contain abnormally large-diameter fibrils. Numerous examples of fibrils with irregular contours or laterally associated fibrils are present (arrows). These structures are indicative of abnormal lateral growth. Bar, 100 nm.

Figure 3.

 

Transmission electron micrographs of corneal sections illustrate structural defects in collagen fibrils indicative of abnormal lateral fusion in the posterior stroma. An overview of a region that contained numerous large-diameter collagen fibrils, many with irregular contours, indicative of abnormal lateral growth (A). The mean diameter of fibrils indicated by the arrows was 87.5 ± 17.7 nm (SD) compared with a normal diameter of 35 nm. A gallery of higher magnification micrographs provides structural details of the abnormal fibrils (B, C, D). Diameters of the fibrils indicated by the arrows were: (B) 135, 91, and 79 nm; (C) 93 and 101 nm; (D) 109 and 63 nm; and (E) 56 nm. (B, C) Especially obvious images indicative of abnormal lateral association and fusion. Corneas were from 7.5-month-old lumican-deficient mice. Bars, 100 nm.

Figure 4.

 

Mean fibril diameter and fibril diameter distribution in anterior versus posterior stroma. The fibril diameter distributions were analyzed in the anterior (A) and posterior stromas (B) in corneas of 7.5-month-old wild-type (lum ⁺/lum ⁺) and lumican-deficient (lum ^tm1sc /lum ^tm1sc ) mice. Masked samples selected randomly from the different regions were analyzed. (A) The anterior stroma of the wild-type and lumican-deficient corneas were nearly identical in mean fibril diameter and distribution, although with a small but reproducible increase in fibril diameter. The fibril diameter range in the anterior stroma was 32 nm and 31 nm for wild-type (minimum, 14 nm; maximum, 46 nm) and mutant (minimum, 17 nm; maximum, 48 nm), respectively. (B) The posterior stroma showed a significant (P < 0.005) increase in mean fibril diameter as well as a shift in the distribution toward larger diameter fibrils. The diameter range was 41 nm (minimum, 22 nm; maximum, 63 nm) and 79 nm (minimum, 21 nm; maximum, 100 nm) for wild-type and mutant posterior stromas, respectively. A population of larger diameter fibrils was observed in the mutant stromas, as seen in the electron micrographs (arrows).

Figure 5.

 

Fibril packing and lamellar organization disrupted in the posterior stroma of lumican-deficient cornea. Transmission electron micrographs taken from approximately 10 μm of the posteriormost stroma from 7.5-month-old wild-type (+/+, A) and lumican-deficient (−/−, B, C) corneas. (A) The lamellar organization of the posterior stroma in lum ⁺/lum ⁺ is regular (bold arrows), with uniformly packed fibrils. (B, C) In contrast, the lamellar architecture of the posteriormost stroma is disrupted in lum ^tm1sc /lum ^tm1sc (−/−) corneas. The fibrils also demonstrate irregular packing and disorganization (∗). Even at this magnification, the large-diameter fibril present in the posterior stroma of the lumican-deficient mice can be seen (B, arrows). Bar, 1 μm.

Figure 6.

 

Increased immunostaining for lumican in the posterior stroma of wild-type control corneas. Corneas were stained with anti-lumican antisera (A, D, E) or secondary antibody only, omitting anti-lumican (negative control, B), or stained with Hoescht to visualize cells (C, F). Three-month-old wild-type cornea showed strong lumican immunostaining in the posterior stroma (S; A, arrows). In 7.5-month-old wild-type stronger lumican expression throughout the stroma reduced the anterior-to-posterior gradient in lumican expression somewhat, although highest lumican expression was still seen in the 20- to 30-μm zone of the posteriormost region (D). Within this zone Descemet’s layer is a thin posteriormost region adjacent to the endothelium (arrows). With the exception of the epithelium, which showed some nonspecific background staining in all samples, lumican-deficient corneas were negative for specific lumican staining, as expected (E). Exposure (integration) times were varied for different samples (A and B: 2 seconds, D: 0.5 seconds, and E: 4 seconds). The 3-month corneas (A, B) were exposed longer than the older corneas (D) to demonstrate the weaker reactivity in the anterior stroma of the younger corneas. The lumican-deficient cornea (E) was exposed eight times longer than the wild-type cornea (D) to demonstrate that the reactivity in the stroma was entirely specific. Bar, 20 μm.