Novel enabling technologies for cataract imaging

The first in vivo detectable sign of diabetic cataractogenesis were hyper-reflective dot-like microlesions [4] that were visualized with a surgical stereomicroscope with dual illumination. The up-to 4 mm depth of focus encompassed the entire NGR lens thickness of ~ 3.4 mm, while the animal’s eye of ~ 6.4 mm transverse radius fitted in the field of view. The microscope dissolved 16 µm at 12.5× magnification [15], which is in the dimension range of the dot-like lesions (Fig. 1A). These distinct and countable microlesions appeared in the lens equatorial regions, with time grew in number, and eventually developed into a full crown (Fig. 1B). These microlesions preceded all cataract types, including PSC, ASC, and cortical cataracts, and were still visible together with early stages of subsequent cataracts (Fig. 1C). However, they were not visible in slit-lamp examination, which is the most commonly used mode of clinical cataract evaluation (Fig. 1D).

Fig. 1

New enabling technologies in cataract imaging. A Schematic of the dual illumination, required for stereo microscopic visualization of the novel hyper-reflective microlesions. B Bright-field image of the first in vivo detectable microlesions in diabetic cataract in the dotted ring stage. C Stereo-microscopy of PSC together with the hyper-reflective microlesions (arrow). D Lack of contrast for the dot-like microlesions in slit lamp biomicroscopy. E Overview of the novel anterior half-globe microscopy technique to image the intact zonules in their native position. External illuminators in combination with epifluorescence microscopy provide a combined 3D image of the ciliary bodies, the zonules, and the migrating cells

To investigate immune cells trafficking to the lens, a novel in situ microscopy technique was developed. In this technique the zonular fibers remained attached to the lens in their original arrangements, while external illuminators in addition to the epifluorescence imaging created a 3D visualization effect (Fig. 1E).

Characterization of the dot-like microlesions

Sections through different frontal planes (Fig. 2A) revealed the microlesions to be in cortical fibers in the front part of the lens and in the subepithelial spaces in the equatorial regions (Fig. 2B–E). There was no protein aggregations or extracellular deposits in these spaces, suggesting accumulation of extracellular fluids. TUNEL assay showed apoptosis in the equatorial epithelial cells and in the nuclei of the elongating fibers in frontal (Fig. 2F, G) and sagittal (Fig. 2H–J) sections, concomitant with subcapsular protrusions, cell proliferation, amorphous material deposits, and vesicular structures inside the lens capsule (Fig. 3A–C).

Fig. 2

Histopathology in the dotted ring stage. A Sagittal lens drawing to illustrate the location of the early dot-like microlesions. B H&E stained frontal section of the microlesions in the cortical region, and C at higher magnification. D The location of microlesions in close proximity to LECs and relative to the nuclei of the elongating fibers, and E at higher magnification. F Overview of TUNEL positive nuclei of the elongating fiber cells in frontal sections, and G at higher magnification. H Overview of sagittal section showing apoptotic nuclei of the elongating fiber cells and LECs, and I and J higher magnifications. TUNEL positive cells co-localized with in vivo detectable hyper-reflective microlesions

Fig. 3

Microstructural lesions in the dotted ring stage and ASC. H&E staining of lenses in the dotted ring stage show A early sub-capsular lesions (black arrows), B amorphous material formation (white arrows) and local LEC proliferation (asterisk), and C vesicular structures inside the lens capsule (black arrowheads). IHC for E-cadherin in lens flatmounts during ASC showed D structural and molecular changes in LECs surrounding ASC lesions (asterisk), E intercellular adhesion defects, F intercellular spaces, and G intracellular vesicle formations. H Cross section of a lens flatmount stained for E-cadherin shows subcapsular amorphous materials. I Intravital microscopy of the anterior part of the lens shows cell contours in areas surrounding the ASC lesion (asterisk), and J extracellular vesicle formations in higher magnifications

E-cadherin, a key adhesion molecule in adherens junctions, is characteristic for the epithelial cell phenotype and its loss associated with EMT [16]. LEC flatmounts of the early ASC cataractous lenses showed incomplete intermembranous E-cadherin expression and intercellular adhesion defects (Fig. 3D–F). Intracellular E-cadherin positive vesicles (Fig. 3G) and accumulation of amorphous materials surrounding the epithelial cells (Fig. 3H) showed progressive loss of adhesion between LECs. Intravital microscopy of ASC cataractous lenses revealed extracellular vesicular structures, suggesting cell dissociation in the early phases of diabetic cataracts (Fig. 3I, J).

Immune cell trafficking in early cataract development

Our new anterior half-globe microscopy technique showed trafficking of the cells attached to the MAGP1 positive zonular fibers (Fig. 4A–C). The migratory cells were CD45 positive leukocytes that likely made their way from the ciliary bodies toward the lens (Fig. 4D–F). In H&E stained sections and in half-globe preparations with additional reflectance illumination, pigmented cells on the zonular fibers became distinguishable (Fig. 4G–I). CD45 positive cells, some of which contained pigments, were located attached to the external surface of lens capsule in the equator region (Fig. 4J–L).

Fig. 4

Immune cell trafficking to the lens in the dotted ring stage of diabetic cataractogenesis. A IHC of MAGP1 to stain the zonular fibers in conventional sagittal section, and B in our new frontal eye globe approach show cells attached to the zonular fibers, and C at higher magnification. D CD45 staining in a conventional sagittal section, and E in our new frontal eye globe preparation, and F at higher magnification show immune cells in the vicinity of the ciliary bodies and zonular fibers. G H&E staining, and H frontal eye globe preparation, and I at higher magnification revealed presence of pigmented cells (arrowheads) on the surface of ciliary bodies and the zonular fibers. J IHC in lens flatmounts demonstrated CD45 positive cells on the lens capsule, K at higher magnification, and L some of which were pigmented (white arrowheads). M Histologic sections of the lens in the dotted ring stage show presence of intra-capsular cells, co-localized with TUNEL positive LECs, and N at higher magnification. Dashed line indicates position of the lens capsule. O Quantification of the migratory cells in the space between the ciliary bodies and the lens in normal, dotted ring stage, and advanced cataractous eyes (n = 3, in each group)

Essential to the lens function is the intactness of the capsule [17]. In the dotted ring stage, we found cells that intruded into the lens capsule. These cells colocalized with TUNEL-positive LECs (Fig. 4M, N). Quantification of the migratory cells showed baseline trafficking in adults with normal lens, which significantly increased in the dotted ring stage, and further significantly increased in the advanced cataractous lens (Fig. 4O).

Alternate routes of immune cell trafficking to the lens

In addition to the constitutive immune cell trafficking along the zonular fibers, immune cells migrated from the retina through the vitreous toward the posterior surface of the lens. However, compared with the immune cells along the zonules, these cells were more sporadic and were only observed in later stages of cataracts (Fig. 5A). Dilation of retinal vessels concurred with cell infiltration into the vitreous (Fig. 5B), while in normal eyes no cells were found (Fig. 5C). In rodents, a pigmented and vascularized pre-optic nerve structure exists, whose function is unknown. In eyes of diabetic animals with advanced cataract, CD68 positive macrophages populated this structure and the retinal surfaces, while cells were not observed in normal eyes. This reveals an as of yet undescribed role for this structure in immune cell trafficking in the eye (Fig. 5D–F).

Fig. 5

The main and alternate routes of leukocyte trafficking to the lens. A DAPI positive nuclei of migrating cells in the posterior part of the eye, and on the surface of the lens (arrowheads). B H&E staining of lenses in advanced cataract show retinal vessel dilation (arrow), and cell infiltration. C H&E staining of the eye of a normal animal shows intact lens and retina. D Retinal flatmount of normal eyes showed no CD68 positive staining in the pre-optic nerve conical structure, whereas E in eyes of diabetic animals with advanced cataract a significant number of CD68 positive cells were observed in the same structure, and F at higher magnification. This suggests an as of yet undescribed role for this pre retinal tissue as a launch pad for immune cells into the posterior chamber of the eye. G Optical coherence tomography (OCT) of the retina of a two month old non diabetic NGR, showing no cells in the vitreous. H Same NGR after LPS treatment, arrow heads, migrating leukocytes through the posterior chamber, and I IHC of the same eye showing the transmigrating leukocytes (arrow heads) in the vitreous body. J Firm adhesion of a leukocyte in a micro vessel of the ciliary body, K transmigration of the leukocyte through the epithelial bilayer of the ciliary body, L its passage through the zonular space, M adhesion to the lens capsule, and N accumulation of several leukocytes at Schlemm’s canal, presumably as prelude to exiting the eye. O Schematic overview of our proposed main and alternate routes of immune cell trafficking through the eye

To map the migration track of these cells, we conducted in vivo imaging experiments. T2D elicits a low-grade inflammatory response, causing migration of few cells at a time. In comparison, acute inflammation through lipopolysaccharide (LPS) causes a pronounced cellular reaction [18], which allowed in vivo tracking. While in a normal NGR no cells were seen in the vitreous cavity (Fig. 5G), a significant number of cells extravasated into the vitreous in the vicinity of the pre-retinal conical structure (Fig. 5H). IHC confirmed that the bright spots were indeed leukocytes that migrated through the vitreous body (Fig. 5I).

With respect to the cells originating from the ciliary bodies, direct dynamic tracking is not possible due to the inaccessibility of the region to light. Serial sections in LPS-treated animals were performed to visualize the trafficking leukocytes at each stage. CD45+ leukocytes firmly adhered to the endothelium of micro vessels in the ciliary bodies (Fig. 5J), transmigrated through the epithelial bilayer surrounding the ciliary bodies (Fig. 5K), and migrated along zonular fibers to the lens (Fig. 5L). Some of the cells were found attaching to the lens capsule (Fig. 5M), while the majority continued their migration through the pupil into the anterior chamber and exited the eye at lymphatic-like structures of the Schlemm’s canal (Fig. 5N). These data map the main and the alternate routes of leukocyte trafficking in the eye (Fig. 5O).

Histopathology of advanced cataract

Frontal sections in advanced cataract showed regular and ordered cortical fibers in the equator region, followed by merged and completely deteriorated fibers toward the lens center (Fig. 6A, B). Sagittal sections along the cortical fibers revealed thin spaces between the neighboring fibers and accumulations of amorphous material therein (Fig. 6C, D). These cortical lesions grew along the lengths of the affected fibers toward the front and the back of the lens. This indicates that the initial cortical microlesions spatially extend and contribute to the developments of PSC and ASC lesions.

Fig. 6

Histopathological characteristics in advanced cataract. A H&E staining of frontal sections shows regular and ordered cortical fibers in the equator region, and B followed by merged and completely deteriorated fibers toward the lens center. C A section along the cortical fiber lengths shows thin spaces between neighboring fibers and accumulated amorphous material therein (arrowheads). D The progression caused widening of the spaces between the fibers and accumulation of more deposits (arrows). E Frontal section of ASC shows local proliferation of cells, while surrounding fibers were normal. F Inline formation of migratory cells in advanced cataractous lens. G Massive disorganization of cortical fibers, sub-epithelial micro-lesions (black arrows), and H sub-capsular amorphous material accumulation were observed (arrowhead). I DAPI positive staining and DIC overlay shows a cell invading into the lens capsule. J Significant capsular defect, surrounded by proliferative cells. K Lens glucose and L GSH concentrations in NGR and STZ-induced diabetic animals. Young non-diabetic NGR (n = 18, 5.6 ± 0.2 m), and diabetic NGR (n = 21, 5.5 ± 0.1 m), aged non-diabetic NGR (n = 12, 13.1 ± 0.2 m) and diabetic NGR (n = 16, 14.2 ± 0.5 m). Long Evans rats, 2 weeks T1D (n = 6), 4 months T1D (n = 4), and age-matched control animals. M Negative correlation between HbA1c and lens GSH in NGR (n = 67, y = − 1.3x + 22.3)

In ASC, actively proliferating cells were surrounded by intact fibers (Fig. 6E). Proliferating LECs migrated from their original location into the lens parenchyma (Fig. 6F). Concurrent with the disorganized cortical fibers, sub-epithelial micro-lesions, sub-capsular amorphous material accumulation, and invasive cells inside the capsule were prominent cytopathologic features (Fig. 6G–I). In its extreme form, the capsular injury showed an irruptive disruption of the ECM, a phenomena that explains why in individuals with cataract lens-proteins is found in the vitreous (Fig. 6J) [19].

To characterize biochemical changes in the lens, glucose and GSH in lenses of young and aged NGRs were compared with T1D diabetic animals at different lengths of hyperglycemia. Both young and aged diabetic NGRs showed significantly higher lens glucose levels than non-diabetic age-matched controls. In 2 weeks hyperglycemic T1D animals lens glucose did not differ from age-matched controls. However, lens glucose was significantly elevated after 4 months of hyperglycemia in T1D animals, compared with age-matched controls (Fig. 6K).

The concentration of GSH, an endogenous antioxidant [20], was significantly lower in the lenses of young and aged diabetic NGRs in comparison with their respective age-matched controls. In T1D animals, the lens GSH concentration was significantly lower after 2 weeks and 4 months of hyperglycemia, compared with age-matched normal controls (Fig. 6L). The GSH concentrations in lenses of NGR negatively correlated with the animals’ HbA1c levels, an indicator for the severity of their diabetes (Fig. 6M).

Characterizations of the glucose metabolism during cataractogenesis

Male and female NGRs in the early stages of cataracts (ASC, PSC, and cortical) showed significantly elevated RBG (Fig. 7A), compared with the dotted ring stage. However, in the dotted ring stage nearly half of all animals—87% of female and 12% of male—showed normal RBG (Fig. 7B). This finding contradicts the osmotic or the glycation hypotheses, whose premise is elevated glucose levels.

Fig. 7

Diabetic cataractogenesis precedes hyperglycemia. A In female and male NGRs RBG significantly increased with transition from the dotted ring stage to the subsequent types of cataracts (ASC, PSC, cortical) (n = 56 animals). B Percentage of female (n = 16) and male (n = 17) NGRs that developed the hyper-reflective microlesions during normoglycemia. C Averaged blood glucose values as measured in OGTT in animals with different stages of cataract formations (n = 28). D AUC0′–240′ for blood glucose was significantly higher in animals with early cataracts and advanced cataracts compared to animals with clear lens. E As opposed to AUC0′–60′, the AUC60′–240′ for blood glucose was significantly higher in the dotted ring stages, when compared with animals with clear lenses. F Average insulin values in OGTT showed distinct insulin response patterns to glucose at different cataract stages. Insulin data showed a biphasic insulin response in the dotted ring stage (n = 28). G Representative graphs from OGTT experiments show blood glucose and plasma insulin measurements in different stages of cataract progression

OGTT and insulin measurements in response to glucose were performed to characterize the systemic metabolism in different stages of cataract formation. The OGTT graphs’ AUC did not differ between the clear lens and the dotted ring stages, while it increased significantly in the early and advanced cataract stages (Fig. 7C, D). While the AUC in the first 60 min between the clear lens and the dotted ring stages were indistinguishable (AUC0′–60′), the latter was significantly higher between 60 and 240 min (AUC60′–240′) (Fig. 7E). The elevation in AUC60′–240′ in the dotted ring stage indicates initial phases of insulin resistance and could become useful in early detection of diabetic complications.

Distinctive insulin response patterns to glucose were found in different stages of cataractogenesis. A biphasic insulin response was observed in most cases during the dotted ring stage (Fig. 7F). Representative graphs illustrate changes in glucose metabolism in individual cases (Fig. 7G). The glucose curves were comparable in the clear lens and the dotted ring stages, while the patterns of the insulin responses differed. In the early cataract stage, despite higher initial insulin values, the animal’s glucose rose to pathologically high levels (> 500 mg/dl) by 60 min. In the advanced stage, the initially high fasted BG (> 300 mg/dl) remained at pathologically high levels throughout the test, while plasma insulin levels failed to rise.

A statistical analysis showed no correlation between the dotted ring stage and the presence or absence of hyperglycemia in NGR. However, all cataract types subsequent to the dotted ring stage, i.e., ASC, PSC and cortical, correlated significantly with hyperglycemia (Table 1).

Table 1 Correlation between cataractogenesis stages with RBG in NGRs