Molecular Vision 2024; 30:478-487
<http://www.molvis.org/molvis/v30/478>
Received 15 June 2023 | Accepted 29 December 2024 | Published 31 December 2024
Molecular Vision 2024; 30:478-487
<http://www.molvis.org/molvis/v30/478>
Received 15 June 2023 |
Accepted 29 December 2024 |
Published 31 December 2024
Sylvain Poinard,1,2 Louise Parveau,1 Gabriel Chapelon,1 Oliver Dorado,1,2 Justin Thomas,1 Zhiguo He,1 Chantal Perrache,1 Alice Ganeau,3 Fabien Forest,1,4 Frédéric Mascarelli,1,5 Philippe Gain,1,2 Gilles Thuret1,2
The first two authors contributed equally to this study.
1Laboratory Biology, Engineering and Imaging for Ophthalmology, BiiO, Health Innovation Campus, Faculty of Medicine, Jean Monnet University, Saint-Etienne, France; 2Ophthalmology Department, University Hospital, Saint-Etienne, France; 3LabTAU, INSERM, Centre Léon Bérard, Université Lyon 1, Univ Lyon, F-69003, Lyon, France; 4Pathology Department, University Hospital, Saint-Etienne, France; 5Cordeliers research center, UMR S1138, Paris Descartes University, Paris, France
Correspondence to: Gilles Thuret, Corneal Graft Biology, Engineering and Imaging Laboratory, EA 2521, SFR143, Faculty of Medicine, Jean Monnet University 10, Rue de la Marandière, 42055 Saint-Etienne Cedex 2, France ; Phone: +33 (0)4 77 12 77 93 Fax: +33 (0)4 77 12 09 95; email:gilles.thuret@univ-st-etienne.fr
Purpose: To date, the assessment of lens epithelial cell viability has been proposed only in cell cultures or isolated capsule models. This study aimed to develop a method for quantifying the viability of epithelial cells on whole ex vivo crystalline lenses by triple labeling Hoechst 33342, ethidium homodimer, and calcein-acetoxymethyl (HEC).
Methods: Two models of induced cell death were used to study the performance and potential applications of the technique. First, ten fresh pairs of six-month-old porcine lenses were retrieved. On one lens of each pair, an easily identifiable localized lesion was induced by a calibrated cryo-application, while the other remained intact. Both lenses of each pair were incubated for 1 h at 20 °C in an HEC mixture. Ten other pairs of lenses were used in the second experiment. On one lens of each pair, a diffuse epithelial lesion was induced by incubation in staurosporine (STS) solution (0.5 µM in CorneaMax) for 24 h at 37 °C. The other lens of each pair was incubated in CorneaMax solution without STS for 24 h at 37 °C. The day after, both lenses of each pair were incubated for 1 h at 20 °C in an HEC mixture. Images were acquired with a macroscope (macro zoom) and analyzed with ImageJ. Calcein-AM and ethidium images were used to calculate the area covered by living epithelial cells. Hoescht images allowed us to count cell nuclei per unit area. Viable epithelial cell density (vECD) was defined as the number of viable cells per unit area. Different strategies were developed to reduce background noise.
Results: There was no interfering lens autofluorescence for the exposure times used. The vECD median was 2,840 cells/mm2 [10th–90th percentiles = 2,479–3,494] for cryo-injured lenses versus 3,364 cells/mm2 [2,919–3,739] for healthy lenses (p = 0.002). The vECD median was 3,804 cells/mm2 [10th–90th percentiles = 2,922–4,862] for lenses treated with STS versus 3,896 cells/mm2 [3,169−4,980] for healthy lenses (p = 0.002).
Conclusions: Thanks to simple sample preparation, triple HEC staining allows fluorescence imaging of a large series of a whole lens to respect the architecture of the epithelium. It will be particularly useful for cytotoxicity studies of new therapies targeting the lens.
Pathological (cataract) or age-related (presbyopia) changes in the crystalline lens of the eye are responsible for the main visual alterations in humans. In situ lens treatments by drugs [1–3] or a physical process [4] represent a significant challenge, given that they would supplant the first surgery in the world (cataract) and be the recourse to methods of compensation for presbyopia (glasses, contact lenses, and corneal surgeries) [5,6]. During the development of these new treatments, it is essential to quantify their toxicity to the lens. The measurement of transparency is essential and must be combined with the measurement of the viability of lens epithelial cells (LEC).
The adult crystalline lens is composed of three main structures: the capsule, the epithelium, and the fiber mass. The monostratified epithelium is present only at the anterior pole and comprises three distinct regions: (1) the center, made of quiescent flat cells, is surrounded by (2) a germinal zone, where cuboid cells divide and migrate toward the equator until they pass it to form (3) a transitional zone of differentiation that, via cellular elongation, is the source of lens fibers throughout life [7]. Thus, the epithelium contributes to the homeostasis of the transparency and deformability functions of the whole lens. It concentrates an intense metabolic activity in comparison to the lens fibers deprived of organelles, particularly the hydroelectrolytic balance by Na+/K+ pumps [8] but also as an emunctory by its rich content in antioxidants [9]. It also participates in the migration and polarization of new elongating lens fibers thanks to connections established at the apical end of epithelial cells [10].
LECs have intense metabolic activity and are essential for lens homeostasis. In particular, their plasma membranes contain transporters that regulate the movement of ions and water necessary to maintain transparency.
Several cytotoxicity assays have already been applied to LECs in culture in vitro (crystal violet [11], 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) [12], cell counting kit-8 (CCK-8) [13]) or on ex vivo lens capsules recovered during cataract surgery (trypan blue [14] and Calcein-AM [15]). The extraction of such a delicate structure cannot be performed without damaging the cells, making it impossible to measure viability accurately, precisely, and reproducibly. Therefore, it is essential to have an in situ method that fully respects the microanatomy of the crystalline lens structure. To our knowledge, no such techniques are available for assessing epithelial viability of the whole lens. One possible reason is the complexity of this model, which has, unlike cell cultures or surgical capsules, a three-dimensional tissue organization with optical properties due to the functions of this organ—transparency and refraction—that disturb its observation.
The live/dead assay combining Hoechst 33342, Ethidium homodimer, and Calcein-AM (HEC) is commonly used in cell biology. It allows the number of living (H+C+), dead (E+), or dying (H+E−C−) cells to be obtained in less than one hour with a standard epifluorescence microscope. Similar to what we adapted for the measurement of corneal endothelial viability in situ [16] [17], we adapted triple HEC staining in situ to the porcine lens to have a reliable (accurate and reproducible) method for quantifying LEC viability while preserving lens integrity.
Twenty pairs of six-month-old pig eyeballs from a local slaughterhouse were used in this study. The time from death to dissection was 2 h. The first step of dissection was the removal of the entire cornea by making a small incision between the cornea and the sclera with a 17-mm diameter trephine and completing the cut with scissors. The iris was then torn with forceps to give free access to the lens. Next, four radial incisions were made on the sclera to facilitate the expulsion of the crystalline lens by pressing on the posterior part of the eyeball. Once the lens was freed, it was immediately placed in 5 ml of an organ culture (OC) medium (CorneaMax, Eurobio, les Ulis, France).
Two models of induced cell death were used to study the performance and potential applications of the technique. The first was a model of localized necrosis involving ten pairs of crystalline lenses to establish proof of concept on easily identified mortality zones. For each pair, one lens was randomly selected to undergo the localized lesion (positive control of epithelial mortality), and the other lens remained intact. The lesion was made on the anterior surface by being toughed for a few seconds with a steel hexagonal rod (to make the mark easily identifiable) that had been previously immersed in liquid nitrogen for one minute. The injured lens was immediately reincubated in OC medium for 30 min at room temperature (RT) to allow for complete cell necrosis.
The second model used staurosporine (STS)—a molecule known to induce apoptosis—to achieve finer and more diffuse mortality than the previous localized acute necrosis. Ten other pairs of crystalline lenses were used. Similarly, for each pair, one lens was randomly selected to undergo the diffuse lesion (positive control of epithelial mortality), and the other lens remained intact. The diffuse lesion was obtained by incubating the anterior face side up in a 0.5 µM STS solution for 24 h at 37 °C. This STS solution was prepared the day before with 2.5 µl of 1 mM STS diluted in 5 mL of CorneaMax and then stored at 4 °C. The healthy lens was incubated in 5 ml of CorneaMax solution without STS for 24 h at 37 °C.
Whatever the model of cell death, both lenses of each pair were placed anterior face side up in a plastic holder and incubated for 1 h at RT (as recommended by the manufacturer) with enough solution to entirely cover the lens’ surfaces (500 µl was sufficient for six-month-old pig lenses thanks to the reduced diameter of the plastic holder). The HEC solution was composed of 10 µM Hoechst 33342 (B2261, Sigma, Saint Quentin Fallavier, France), 2 mM Ethidium homodimer-1 (FP-AT758A, Interchim, Montlucon, France), and 4 mM Calcein-AM (FP-FI9820, Interchim, Montlucon, France) in Optimem (11058021, Gibco, Thermo Fisher, New York, NY) [16]. Hoechst 33342 stained nucleic acids by binding to the minor groove of double-stranded DNA. Due to its lipophilicity, it has the ability to cross intact cell membranes. Once bound to DNA, Hoechst 33342 was thirty to forty times more fluorescent than in its unbound form. It was excited by ultraviolet light at about 350 nm and emitted blue/cyan fluorescent light, with a peak at 461 nm. Ethidium homodimer-1, a DNA intercalator, has the ability to penetrate cells with permeable membranes and thus stain the nuclei of dead cells. Once bound to DNA, ethidium homodimer-1 was thirty times more fluorescent. Its maximum fluorescence was observed at 617 nm for a maximum excitation wavelength of 528 nm. Calcein-AM is a nonfluorescent molecule that crosses the intact plasma membrane. It is converted to calcein by ubiquitous intracellular esterases. Calcein was fluorescent, with maximum excitation at 495 nm and an emission peak at 515 nm. Calcein, being hydrophilic, was retained inside the living cells. Thanks to this triple staining, we were able to obtain results expressing the epithelial cell density and the viable epithelial cell density (vECD). Hoechst allows for labeling each nucleus as either dead or alive to obtain an accurate nucleus count. Ethidium and calcein were used to calculate viable and dead areas. After incubation in this HEC solution, the crystalline lenses were rinsed in a balanced salt solution (BSS) (BSS Plus®, Alcon, Rueil-Malmaison, France) and immediately observed.
Before carrying out the mortality models with cryo-application and STS, we ensured that staining properly marked epithelial cells, not lens fiber cells. To do this, we stained a cryo-injured lens and took microscope images of the lens before and after cutting it on the equatorial plane with a vibratome. After cutting, we observed no fluorescence in the posterior halve of the lens, except for a small band at the edge of the lens that had not been cut, corresponding to the remaining equatorial epithelial cells (Appendix 1).
For image acquisition, the anterior surface of the lenses was positioned upward toward the macroscope placed above the samples. They were immersed in BSS to reduce the difference in refractive index between the air and the sample and thus avoid reflections from the curved anterior capsular surface. The reflections of the posterior pole of the lens after light scattering within its transparent structure were minimized by this technique and by the use of a container with a matte black bottom. The lens was immobilized by positioning its posterior surface on a drop of viscoelastic substance (Viscoat, Alcon, Rueil Malmaison, France) placed at the bottom of the container before covering it with BSS.
Two observation methods were then used: (1) calcein (isothiocyanate de fluorescéine [FITC]) and ethidium (cyanine 3 [CY3]) fluorescence were imaged at low magnification (×1 objective) with a macroscope (MVX10, Olympus, Tokyo, Japan) equipped with a light source (X-Cite 120PC Q, Mississauga, Canada) and CellSens software (Olympus, Hamburg, Germany). The 5,700 × 3,600 pixel images for 17,860 µm and 11,163 µm were saved in a tagged image format file. These two images provided the areas covered by live and dead cells, respectively, for which a cell-level resolution was not required. (2) The density of LECs (in cells/mm2) was measured on an image mosaic (Multiple Image Alignment) on 4',6-diamidino-2-phénylindole (DAPI) images at higher magnification (×10 objective) and on a flat surface to avoid defocus and parallax errors. For this purpose, the lens was slightly flattened between two microscope slides spaced by two 6-mm-thick wedges. Images were acquired with a fluorescence microscope (IX81, U-TBI90 Olympus, Tokyo, Japan) equipped with a light source (X-Cite 120PC Q, Mississauga, Canada) and CellSens software (Olympus, Hamburg, Germany).
Image analysis was conducted in two main steps using two different software packages. The first step used a custom macro in Fiji software to standardize the ECD counting zones. For the second step, the CorneaJ plugin allowed us to count nuclei on the isolated Hoechst images and quantify viability (on the calcein+ area) and mortality (on the ethidium+ area).
In the first step of the macro, a region of interest (ROI) of 8 mm in diameter (50.3 mm2) was centered on the lens axis and delimited by the flattened, focused area on the Hoechst image. In this ROI, five square areas of 200 × 200 µm were isolated: one in the center and four in the periphery to account for the centroperipheral gradient of ECD already reported in the literature [18]. For injured lenses, the lesion zone was avoided for counting. These five zones were chosen on the injured lens and then reported at the exact same location on the control lens. In the second step, the CorneaJ plugin was used to count the number of nuclei per unit area (cells/mm2) in each of the five selected zones of the Hoechst images. Images were 2,048 × 2,048 pixels with a pixel ratio of 1.5385 px/µm. A manual threshold was applied to isolate each nucleus. The largest possible ROI was selected to increase counting reliability (Figure 1). In the second step, viability was analyzed using the CorneaJ plugin [17]. The images were calibrated with a pixel ratio of 0.3225 px/µm for an image resolution of 5,760 × 3,240 pixels. A manual threshold was applied to binarize the image and select the viable area only (calcein+) in the same 8-mm diameter ROI. The same procedure was repeated on ethidium+ images to analyze the percentage of surfaces covered by dead cells. The combination of these measurements (viability and epithelial cell density) allowed us to calculate the number of viable epithelial cells per unit area (vECD = ECD × percent viable area). These unique data, which were always measured in the 8-mm center of the anterior capsule, allowed for easy comparison of the different lenses.
Given the sample size, data were expressed as median (10th–90th percentile). Quantitative data (ECD and vECD) were compared using the nonparametric Wilcoxon Mann–Whitney test. Spearman’s R correlation coefficient was calculated for the ECD of the paired lenses. A p value < 0.05 was considered significant.
Triple HEC staining of the entire lens allowed differentiation between areas covered with living cells, dead cells, and those without cells (Figure 2). In this series of ten cryo-injured lenses, the median percentage of the calcein+ area in the 8-mm diameter ROI was 90% (86%–92%), and for the ethidium+ surface, it was 11% (8%–14%). The complementarity between the two measurements was therefore considered excellent. The images also revealed two phenomena: (1) zonule and ciliary body residues strongly fixed ethidium and formed a fluorescent ring around the lens, and (2) the fluorescence intensity of calcein was more important at the periphery of the lens.
The ECD (H+ nuclei per unit area) was calculated by considering many cells—1,748 (685–2,345) cells per lens—for the whole series. The ECD, measured outside the destroyed areas in the cryo-injured series, was 3,206 (3,024–3,862) cells/mm2 for the treated lenses versus 3,364 (2,935–3,926) cells/mm2 for the control group. These were established in the same counting areas thanks to standardized software (p = 0.912; Figure 1). The ECD of the two lenses of the same animal were therefore strongly correlated and indicated a certain reproducibility of the method (Figure 3).
In the study involving the localized necrosis lesion, the median percentage of viable surface areas was 90% (86%–92%) for the treated group, whereas it was 99% (93%–100%) for the control group, reflecting the ability of the staining to show small epithelial lesions caused by lens dissection. The vECD was 2,840 (2,479–3,494) cells/mm2 for the treated group versus 3,364 (2,919–3,739) cells/mm2 for the control group (p = 0.002; Figure 4).
In the study involving the chemical toxicity caused by STS, a subtle increase in mortality was induced to confirm the sensitivity of the technique: the vECD was 3,804 (2,922–4,862) cells/mm2 for the treated group versus 3,896 (3,169–4,980) cells/mm2 for the control group (p = 0.002; Figure 5).
Given the crucial role of the epithelium in eye lens homeostasis, a reliable measurement of LEC viability/mortality is essential to quantify the toxicity of treatments or processes applied to this organ. We proposed an adaptation of the triple HEC staining that we originally described for the corneal endothelium. For the cornea, it was a question of improving the inaccuracy that existed in the measurement of ECD by optical microscopy without staining to carry out a measurement on the whole endothelium and to integrate the percentage of viable cells into the count. This technique has since been adopted by many authors and has become the gold standard for preclinical experiments [19–22].
Most articles reporting on LEC viability studies use cell culture models with standard colorimetric techniques. Although essential for preliminary exploration, these simplified models do not always reflect the physiologic environment of the epithelium. Moreover, colorimetric measurement requires the use of spectrometers whose medium must be homogeneous within absorbance cuvettes to satisfy the proportionality principle of the Beer–Lambert law. The assessment of the viability of a given cell at the tissue level is complicated by the interaction of the marker with other components of the tissue, including the delivery and observation of the marker. Trypan blue is one of the oldest and best-known vital dyes and can only enter a cell with compromised plasma membrane integrity. However, a parallel technique is required to count all cells to establish the proportion of dead cells. Moreover, trypan blue has an affinity for collagens, and this property is exploited to increase the visualization of membranes before peeling them in endothelial corneal grafts (Descemet stripping) and cataract surgery (Capsulorhexis). This double cell and membrane affinity hinders the measurement of LEC viability on the whole lens by decreasing the contrast between cells and the extracellular matrix. Furthermore, the refractive and scattering properties of the lens prevent sufficient backlighting of the trypan blue-labeled LEC when observed under a phase contrast microscope. For all these reasons, the use of fluorescence markers is more suitable for this application. While immunohistochemistry has been performed on whole mouse lenses using a confocal microscope [23], proof of concept for measuring LEC viability has only been established on isolated capsules [24].
We encountered several difficulties in adapting the HEC technique to the whole lens. First, the transparent and refractive nature of the tissue is the source of various optical artifacts. The lens capsule reflects light and requires the removal of all external light sources. Reflections are also created by the fluorescence of the markers themselves, which diffuses in all directions, particularly toward the posterior surface. The hyperfluorescence of calcein at the periphery of the lens compared to the center can be explained by these optical properties (internal reflection of light), by its observation in front of this curved zone, and by the presence of a greater density of cells in the germinal zone with strong metabolic activity.
To minimize these artifacts, we immersed the lens in a liquid (BSS) to reduce the difference in refractive indices between the air and the tissue. Immersion also eliminates the risk of capsules drying out and forming autofluorescent asperities. The second approach is to place the lens in a container with matte-black walls. The lens is simply stabilized at the bottom of the box by a viscoelastic substance. By keeping it at a distance from the bottom of the container, we also avoided most reflections caused by contact between the posterior pole of the lens and the container wall.
Second, the curvature of the anterior surface prevents cell counting in Hoescht in a single focal plane, even with a macroscope with a large depth of field. For this step, we slightly flattened the lens between the two glass slides. We standardized this flattening using wedges adapted for six-month-old porcine lenses. Finally, we counted the cells using a Fiji macro to ensure that the procedure was ergonomic, standardized, and reliable.
Two models of cell death were employed in this study. The first involved localized frostbite, which allowed for the easy identification of mortality zones and confirmed the complementarity of the markers. Additionally, due to its massive (large area) and brutal nature, it implies a mode of mortality close to acute necrosis. This approach could be used in studies involving acute mechanical stress, such as the use of physical agents for presbyopia treatment, to assess their safety [25–27].
The second model evaluated the technique by exploring the apoptotic aspect of a pharmacological application using a validated molecule—STS [28]. A small variation in mortality was detectable, which could make the technique useful for preclinical studies assessing the toxicity of drug candidates [29–32].
In summary, the use of triple HEC staining adapted to the whole lens provides an unrivalled measure of pan-epithelial viability. It allows easy processing of a large number of samples without the need for tedious capsule dissection, which destroys the lens’s architecture. Moreover, even if the volume of the HEC solution is greater than that required for a flat sample, the cost of reagents has been estimated at about 50 cents per lens.
Triple HEC staining provides a tool for measuring the viability of the epithelium on whole eye lenses. Due to its simplicity and low cost, this pan-epithelial evaluation established on a porcine model allows for the treatment of a significant number of samples, which is a prerequisite for the development of new therapeutics applied to the lens.
Supported by The National Agency for the Safety of Medicines and Health Products 301 (L’Agence nationale de sécurité du médicament et des produits de santé or ANSM), appel à 302 projet recherche 2012, BANCO project, the Fondation de l’Avenir pour la recherche médicale 303 appliquée, APRM16008 project and the French National Research Agency. ANR 304 Presbinnov-US ANR21CE190059. Disclosure: no proprietary or commercial interest in 305 any materials discussed in this article (None). Grants: Supported by The National Agency for 306 Research ANR PRESBINNOV-US (ANR21CE190059), appel à projet recherche 2012, 307 BANCO project, and by the Fondation de l’Avenir pour le recherche médicale appliquée, AP308 RAM16008 project.Translational Relevance: a reliable measurement of the ex vivo lens 309 epithelial viability will accelerate the development of drugs or processes to modify the lens.