Collagen matrix perturbations in corneal stroma of Ossabaw mini pigs with type 2 diabetes

Animals

All experimental procedures were approved by the Institutional Animal Care and Use Committee, University of Missouri Columbia, and adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Adolescent female and male Ossabaw mini pigs (RRID:NSRRC:0008) were housed at the Animal Science Research Center, the University of Missouri, Columbia, MO, in a temperature-controlled environment with 12 h light and 12 h dark cycle. Following weaning, all pigs were fed on a standard commercially available chow diet (catalog no. 5L80, Lab Diet; 3.03 kcal/g, 71% from carbohydrate, 18.5% from protein, and 10.5% from fat) until 3.5 months of age. Twelve male and female mini pigs were divided into two groups equally and fed on SD and WD for ten weeks. SD mini pigs were kept on the standard chow, while pigs in the WD group were fed a high-fat/high-fructose corn syrup/high-calorie diet (5B4L, Lab Diet; 4.14 kcal/g, 40.8% from carbohydrate, 16.2% from protein, and 43% from fat). At six months of age, following a 20-h overnight fast, pigs were anesthetized (5 mg/kg of Telazol and 2.25 mg/ kg of xylazine mixture) and weighed (for body mass determination), blood was collected, and mini pigs were subsequently euthanized for tissue collection as described earlier [33]. Both retinal and corneal tissues were collected post-euthanasia in the study.

Ossabaw pig tissue collection

Following euthanasia, eyes were enucleated with surgical forceps and Westcott scissors. An incision was made at the ora ciliaris, approximately 5 mm posterior to the corneal limbus, using a number of 11 scalpel blade. A circumferential cut was made to remove the anterior ocular tissues, including the ciliary body and lens. The ciliary body, lens, and iris were removed from the cornea using surgical forceps and Westcott scissors under a dissecting surgical microscope (Leica Wild M690, Leica Microsystems Inc., Buffalo Grove, IL). The corneal tissues were placed in 1.5 ml Eppendorf tubes and stored at −80°C. Later, corneas were cut into three equal parts: one part was used for histology studies, and the other two parts were used for molecular studies and transmission electron microscopy (TEM) analysis. For histology, corneal tissue was immediately placed into molds containing an optimal cutting temperature (OCT) compound and snap-frozen in a container of 2-methyl butane immersed in liquid nitrogen. Frozen tissues were maintained at −80°C until sectioning and further evaluation. Tissues were sectioned at 8 μm thickness, mounted on microscopic glass slides (SuperFrost Plus, Fisher Scientific), and slides were preserved at −80°C for subsequent analysis.

Hematoxylin and Eosin (H&E) staining

H&E staining was performed following the standard technique [35]. Briefly, cryo-frozen corneal sections were incubated for 15–20 min at room temperature, washed in PBS for 10 min, and dipped in hematoxylin for 5 min, followed by a rinse in running DI water, one dip in 1% acid-alcohol, and 1 min in 0.3% ammonia water. The slides were incubated in 95% alcohol, followed by eosin. The tissues were then dehydrated in absolute alcohol, cleared in CitriSolv solution (Decon Laboratories, King of Prussia, PA), and finally mounted with cytoseal (Richard-Allan Scientific, Kalamazoo, MI). The images of H&E-stained tissues were captured with a bright-field microscope (Leica) equipped with a digital camera and imaging software (SpotCamRT KE; Diagnostic Instruments, Sterling Heights, MI).

Masson trichrome and picrosirus red staining

Eight-micron thick Ossabaw pig corneal tissue sections from the WD and SD groups were subjected to Masson’s trichrome staining to evaluate the alterations in collagen, a primary component of extracellular matrix (ECM). Masson’s trichrome and Picrosirus Red staining was performed by the Veterinary Medical Diagnostic Laboratory at the University of Columbia, MO, as described earlier [35].

Collagen I Immunohistochemistry

Eight-micron thick corneal sections were prepared, postfixed at room temperature for 10 min, and blocked with 5% donkey serum for 1 h at room temperature. Immunostaining was performed using collagen I (COL1) (1:200; Abcam), kept overnight at 4°C, and secondary stained with Alexa-Fluor 594 antibodies (1:500; Life Technologies) for 4 h. A drop of DAPI antifade Vectashield medium (H1200) was applied, and sections were mounted with premier coverslips (Thermo Fisher). (DAPI; Vector Laboratories) The stained sections were viewed and photographed with a fluorescence microscope (Leica DM 4000B, Leica Microsystems Inc., Buffalo Grove, IL) equipped with a digital camera (SpotCam RT KE, Diagnostic Instruments Inc., Sterling Heights, MI). Negative control samples were stained by excluding primary or secondary antibody in serially sectioned slides.

TUNEL assay

Keratocyte apoptosis in the corneas of the WD- and SD-fed mini pigs was determined by TUNEL assay (ApopTag; Millipore, Temecula, CA). Corneal sections were fixed in 1% PFA for 10 min, and a TUNEL assay was performed following the manufacturer’s instructions, including suitable positive and negative controls. Rhodamine-conjugated apoptotic cells (red) and 40,6-diamidine-20 -phenylindole dihydrochloride (DAPI)–stained nuclei (blue) were viewed and photographed with a fluorescence microscope (Leica) fitted with a digital camera system (SpotCamRT KE; Diagnostic Instruments, Sterling Heights, MI). DAPI-stained nuclei and TUNEL-positive cells in untreated and treated tissues were quantified at 100x magnification in six randomly selected non-overlapping areas [36,37]. Negative control samples were stained by excluding the Tdt enzyme or Anti-Digoxigenin Conjugate in serially sectioned slides.

Transmission electron microscopy (TEM)

The ciliary body, iris, and lens were removed as described. One-fourth of the cornea was placed in one milliliter of TEM primary fixative (2% paraformaldehyde, 2% glutaraldehyde in 100 mM sodium cacodylate buffer; pH 7.35). Tissues were then transitioned into acetone, infiltrated with Epon resin (250 W for 3 min), and polymerized at 60°C overnight. Corneas were cut transversely with an ultramicrotome to obtain 2 µm thick sections for histologic examination. The cornea was further trimmed to a length of 500 µm and cut with a diamond knife to yield 75 nm sections for TEM. Images were acquired with a JEOL JEM 1400 transmission electron microscope (JEOL USA Inc., Peabody, MA) at 80 kV on a Gatan Ultrascan 1000 CCD, at 20000x magnification.

For automated quantification of corneal collagen fibrils in TEM images, an in-house image processing and analysis software was developed as described in the previous study [38]. Briefly, the algorithm comprised three main modules, (i) fibril detection, (ii) shape analysis and cluster decomposition, and (iii) size and interfibrillar distance analysis. A multi-scale Hessian matrix was used to detect collagen fibrils. Hessian matrix (Equation 1) described the second-order structure of local intensity variations around each point of the image L (x, y), and eigenvalues ƛ_1,2 (Equation 2) of the Hessian matrix was used to detect blob-like, or ridge-like structures. Table 1 shows possible local orientation patterns based on the eigenvalues of the Hessian matrix.

H e s s i a n σ ( X , Y ) = L x x ( X , Y )  L x y ( X , Y ) L x y ( X , Y )  L y y ( X , Y ) λ 1 , 2 = 1 2 ( L x x + L y y ± ( L x x + L y y ) 2 + ( 2 L x y ) 2 )

Hessian matrix was computed by convolving the image with derivatives of the Gaussian kernel, where scale s represented the standard deviation of the Gaussian kernel and controlled the radius of the detected structures. A coarse fibril mask was produced by computing Hessian matrix and thresholding ƛ₁ (Hessian) as below (Equation 3):

M a s k f i b r i l ( X , Y ) = 1 λ ( H e s s i a n ( x , y ) < ε 0  otherwise

The quantification of morphology and spatial organization of fibrils requires accurate identification and localization of individual fibrils. Hessian-based detection efficiently segmented regions occupied by fibrils from the background but failed to separate some neighboring fibrils and merge them into clusters. To identify individual fibrils, shape analysis and cluster decomposition module were developed based on previously reported modules [39,40]. Specifically, the first connected component labeling was applied to the detection mask and disconnected blobs. Then, to each detected blob B_i, an ellipse E_i was fitted. Blob and ellipse areas: area (B_i), area(E_i), and ratio r=area(E_i) - area(B_i))/area(B_i) were then computed. Each detected blob Bi was classified into one of the three classes (spurious detection, single fibril, or fibril cluster) using size and shape indices and their means and variances over the image. Blobs classified as spurious detection were removed; blobs classified as single fibril were kept intact. The marker-controlled watershed transformation was used to decompose fibril clusters into individual fibrils [41]. Regional maxima of distance transform were used as markers. To suppress spurious regional maxima and prevent over-segmentation, H-maxima transform was applied to the distance transform before detecting regional maxima [42].

The module then computed parameters concerning fibril morphology and spacing, such as fibril radius, fibril area fraction, and interfibrillar distance. From the refined segmentation, fibril centroids were computed. Delaunay triangulation and vertex coloring were applied to the located centroids, and a colored neighborhood graph was generated, as previously described [39,43]. In the neighborhood graph, nodes corresponded to individual fibrils, and edges linked immediate neighbors. For each node in this neighborhood graph, two specific neighbors were identified: the nearest immediate neighbor (red edges) and the farthest immediate neighbor (blue edges). A second graph (nearest/farthest neighborhood graph) was constructed using only these specific links. Fibril-to-fibril interactions were assessed, and the spatial organization of collagen fibrils were described using these two neighborhood graphs. Various descriptors from these graphs, such as mean and standard deviation, were then calculated as described previously [38].

Quantitative reverse transcriptase PCR (qRT-PCR)

For molecular studies, corneal tissues were minced in a tissue lyser (TissueLyser LT, Qiagen) in RLT buffer (Qiagen, Valencia, CA), and total RNA was isolated using the RNeasy kit (Qiagen, Valencia, CA) following the manufacturer’s instructions. The reverse transcriptase enzyme kit was used to synthesize first-strand cDNA (Promega, Madison, WI) [44]. The One Step Plus Real-Time PCR system (Applied Biosystems, Carlsbad, CA) was used for quantitative PCR (qPCR). A 20μl reaction mixture containing 2μl cDNA, 2μl forward and reverse primers (200nM each), and 10μl of 2X All-in-One PowerUp SYBR green master mix (Applied Biosystems) was run at a universal cycle (95°C for 10 min, 40 cycles at 95°C for 15 s, and 60°C for 60 s), as previously reported [44]. Forward and reverse gene-specific primer sequences are shown in Table 2. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as an endogenous reference gene. The threshold cycle (Ct) was used to detect the increase in the signal associated with the exponential growth of PCR products during the log-linear phase. ΔCt for each sample was calculated by subtracting the Ct of the target gene from that of the Ct of the endogenous reference gene, and ΔΔCT was calculated by subtracting the ΔCt of the test sample from that of the ΔCt of the control sample. The relative mRNA expression was calculated using the 2-ΔΔCt method and reported as a relative fold change over the corresponding control values. The amplification efficiency for the qRT-PCR was similar for all templates used, and the difference between linear slopes was below 0.1. The qPCR was performed in triplicate for each sample, and a minimum of three independent experiments was conducted.

Statistical analysis

The GraphPad Prism 8.2.1 (GraphPad Software, La Jolla, CA) software was used for statistical analysis. Each experiment was conducted independently in triplicate, and the values were expressed as mean ± SEM. For statistical analysis, the Student t-test and two-way ANOVA with Bonferroni post hoc test were used. The value of p ≤ 0.05 was considered as the significance level.

Ossabaw mini pigs fed on WD exhibits signs of T2DM

Ossabaw mini pigs fed on WD showed clinical signs of metabolic syndrome and hyperglycemia, demonstrating characteristic features of T2DM. Ossabaw mini pigs showed a significant increase in body mass, cholesterol, dyslipidemia, and significantly high plasma glucose after 10 weeks on WD compared to the SD-fed pigs [33]. Additionally, WD-fed pigs showed elevated insulin resistance levels based on the Homeostatic Model Assessment of Insulin Resistance (HOMA-IR), a clinical indicator of insulin resistance described in our previous study [33].

Ossabaw mini pigs fed on WD displays increased activation and apoptosis of keratocytes

H&E staining in WD pig corneas showed intensified nuclei staining in the stroma (Figure 1B) compared to SD pigs (Figure 1A). TUNEL assay performed in the Ossabaw mini pigs suggested elevated apoptotic positive cells = red in WD corneal stroma (Figure 2B) compared to the SD group (Figure 2A). Total number of tunnel positive cells per field were considered for objective comparison between the two groups.

Ossabaw pigs fed on WD shows irregularity in the corneal stroma

We used Mason Trichome, a tricolor stain, which showed increased collagen (blue), reduced cytoplasm staining (pink/red), and increased cell nuclei (black) in WD pigs (Figure 3B) compared to SD pigs (Figure 3A). Immunohistochemistry staining of T2DM pig corneas with COL 1 displayed an irregular distribution of collagen fibrils in the WD pigs (Figure 4B) compared to SD pigs (Figure 4A). This observation was further confirmed with the use of Picrosirus Red staining, where Ossabaw pigs on a WD increased collagen staining (red) in the stroma (Figure 5B) compared to the SD control group (Figure 5A).

Ossabaw mini pigs fed on WD shows ultrastructural defects in stromal collagen matrix

Corneas were analyzed ultrastructurally for the evaluation of stromal collagen fibril arrangements (Figure 6 and Figure 7). TEM showed marked alterations in stromal collagen fibril size and fibril organization. We used computer-generated collagen ellipses of transverse corneal TEM images (Figure 6). A significant change was noted in areas measured regarding the fibrils’ frequency (p = 0.0034). Ossabaw mini pigs on WD had smaller-sized collagen fibers compared to the pigs on SD. SD mini pigs had a range of collagen fibers with two average sizes of approximately 300 pixels and 900 pixels. In contrast, WD mini pigs had a lower number of small fibers, and the average size of larger fibers decreased from 900 fibers to approx. 800. Adding to size differences, WD pigs had differences in the interfibrillar distance (IFD). Figure 7 describes a fibril graph connecting individual collagen fibrils with their neighbors. Further analysis on fibril distances was performed by mapping minimum (red lines) and maximum (blue lines) IFD. WD mini pigs showed statistically significant changes in minimum (p = 0.0212) and maximum (p = 0.0018) IFD compared to the SD group, indicating a perturbation in collagen distribution corneal stroma.

Ossabaw mini pigs fed on WD depicts altered gene expression related to collagen and fibrosis in the corneal stroma

Collagen-related gene expression also showed a significant decrease in collagen IV (COL4) expression and increased matrix metalloproteinase 9 (MMP9) expression in the WD diet-fed Ossabaw mini pigs compared to the SD diet-fed pigs. In contrast, Collagen III (COL3), MMP1, and Tissue Inhibitor of Metalloproteinase 1 (TIMP1) showed no change in the gene transcripts (Figure 8A). We also studied the expression of fibrosis-related genes in the pig corneal stroma. Alpha-Smooth muscle actin (αSMA) and fibronectin (FN) showed significantly elevated expression, whereas TGFβ1 was insignificant in the WD group when compared to the SD group (Figure 8B).