Abstract
purpose. To examine ultrastructurally the composition of major extracellular matrix (ECM) components and the distribution of myocilin in the trabecular lamellae of corneoscleral (CS) meshwork in normal human eyes. The codistribution of myocilin with ECM components was also investigated.
methods. Postembedding immunoelectron microscopic studies were performed with antibodies against myocilin and other ECM components, including fibronectin, laminin, vitronectin, tenascin, elastin, fibrillin-1, microfibril-associated glycoprotein (MAGP)-1, decorin, versican, hyaluronic acid, and five types of collagen (I, III, IV, V, and VI). Double labeling of myocilin with other ECM components was performed with different sized gold particles.
results. In the trabecular beams of CS meshwork, fibronectin, laminin, and collagen type IV were associated with basement membranes, whereas elastin was specifically localized to the core of elastic-like fibers. Several types of collagens, glycoproteins, proteoglycans, and hyaluronic acid were detected both in the collagen fibers and ground substances. Myocilin predominantly localized in the long-spacing collagens and sheath materials surrounding elastic-like fibers, codistributed with fibronectin, fibrillin-1, MAGP-1, decorin, and type VI collagen.
conclusions. This study illustrated the composition of ECM materials in the trabecular lamellae of CS meshwork. Myocilin was specifically localized to long-spacing collagens and the surrounding sheath of elastic-like fibers interacting with microfibril-associated elements where changes have been documented to occur in glaucomatous and aging eyes.
Myocilin,
1 also known as trabecular meshwork-inducible glucocorticoid response (TIGR), was originally cloned from cultured human trabecular meshwork (TM) cells as a protein upregulated by dexamethasone treatment.
2 3 This gene has been directly linked
4 to both juvenile- and adult-onset primary open-angle glaucoma (POAG). Approximately 3% to 4% of patients with POAG have been found to have mutations in this gene, irrespective of region or race.
5 6 7
In humans, myocilin is expressed in several ocular and nonocular tissues.
8 The expression level of myocilin in the TM appears to be considerably higher than in other ocular tissues. In situ hybridization experiments also indicate that the mRNA levels are similar in the uveal, corneoscleral (CS), and juxtacanalicular (JCT) regions of TM tissues.
9 10 11 12 13 The physiologic functions of myocilin and its precise roles in the pathogenesis of glaucoma nevertheless remain largely a mystery.
8 We have demonstrated by immunoelectron microscopy that myocilin is localized both intracellularly and extracellularly to multiple sites in normal human TM tissues and in cultured cells.
14 Intracellularly, myocilin is associated with mitochondria, vesicles, centrosomes, and cytoplasmic filaments, including actin stress fibers and intermediate filaments. Extracellularly, it is localized in TM tissues in association with the extracellular matrix (ECM).
In a subsequent investigation,
15 we examined the extracellular localization of myocilin in the JCT connective tissue of TM. Myocilin was demonstrated to localize predominantly to the microfibrillar architectures of sheath-derived (SD) plaque materials, within the clusters of the banded material surrounding the plaques. This distribution pattern is intriguing, because the microfibrillar structure is the most prominent ECM component in the region and abnormal accumulation of SD plaques has been established as a characteristic pathologic change observed in the JCT of patients with POAG.
16 17 18 19 20 21 22
The CS meshwork portion of TM tissues extends approximately 100 μm in the direction of the aqueous flow and consists of interconnecting sheets of trabecular beams that contain lamellae of connective tissue elements.
22 23 Cells in the CS meshwork line the trabecular beams. This contrasts with that in the JCT region, where cells reside relatively freely and are embedded in connective tissues. The JCT/Schlemm’s canal area is believed to be the main site of resistance of the aqueous outflow. The TM cells covering the corneoscleral beams nevertheless are also likely to have their roles in maintenance of the normal outflow. For instance, TM cells are known to engulf debris in a self-cleaning manner,
24 so that the pathways remain free for the fluid to flow through the meshwork.
Underneath the TM cells in the CS meshwork, there are basement membranes. The beams or trabecular cores are made up essentially of collagen fibers embedded in ground substances as a matrix, and elastic-like fibers as a plexiform framework. In trabecular lamellae, the accumulation of “long-spacing collagens” and thickening of basement membranes are well-documented changes in aged eyes.
25 Long-spacing collagen is also observed in the eyes of patients with POAG.
16 It is characterized as a cross-banded structure with approximately 100 to 120 nm periodicity,
26 which is longer than that (52–62 nm) of collagen types I and III. The exact nature of this structure has yet to be clearly defined.
In this study, we continued our efforts, analyzing systematically the ECM composition of the basement membrane, the trabecular core and the long-spacing collagens in the CS meshwork of normal human eyes. A postembedding colloidal gold immunoelectron microscopy (immuno-EM) method was used. The TM sections were immunostained with antibodies specific for major ECM components of TM including fibronectin, laminin, vitronectin, tenascin, elastin, fibrillin-1, microfibril-associated glycoprotein (MAGP)-1, decorin, versican, hyaluronic acid, and five types of collagen (I, III, IV, V, and VI).
The extracellular localization of myocilin in this region was also investigated. In addition, double labeling was performed to determine whether, and which, ECM components colocalize with myocilin in the trabecular lamellae of CS meshwork.
Six normal human donor eyes (donor ages, 39, 47, 48, 58, 72, and 74 years) with no history of glaucoma or other eye diseases were obtained from the Illinois Eye Bank at Chicago within 24 hours of death. The procurement of tissues was approved by the Institutional Review Board at the University of Illinois at Chicago and complied with the Declaration of Helsinki. TM tissues isolated were fixed for 3 hours in 4% paraformaldehyde and 0.1% glutaraldehyde in 0.1 M phosphate buffer (pH 7.4). The specimens were dehydrated at −20°C through a graded series of N,N-dimethylformamide and were finally immersed in a glycol methacrylate embedding medium. They were subsequently polymerized by ultraviolet irradiation at −20°C for 48 hours. Ultrathin 80-nm serial sections were obtained through the entire layers of TM, and were mounted on 150-mesh nickel grids.
To minimize nonspecific binding, sections were blocked at room temperature (RT) in 1% bovine serum albumin in phosphate-buffered saline (PBS) for 15 minutes. The grids were then incubated at RT with a specific primary antibody for 3 hours. The primary antibodies were rabbit anti-human fibronectin (1:20 in the blocking buffer; ICN Biochemicals, Irvine, CA), laminin (1:15; Sigma-Aldrich, St. Louis, MO), tenascin (1:20; Invitrogen-Life Technologies, Gaithersburg, MD), elastin (1:150; Elastin Products Co., Owensville, MO), fibrillin-1 (1:150; Elastin Products Co.), MAGP-1 (1:200; Elastin Products Co.), and collagen types I (1:20; Chemicon, Temecula, CA), III (1:40, Chemicon), and IV (1:20; Collaborative Research, Bedford, MA); mouse anti-human vitronectin (1:20; Invitrogen-Life Technologies), versican (1:15; Seikagaku Corp., Falmouth, MA), and collagen types V and VI (1:15; Chemicon); and sheep anti-human decorin (1:500; United States Biological, Swampscott, MA). Several dilutions of the antibodies were tested, and the optimal one was chosen for the study. An affinity-purified polyclonal antibody anti-myocilin-33 was also used as a primary antibody at a dilution of 1:200. The development and specificity of this antibody were as previously described.
14 15
After the primary antibody incubation, the grids were rinsed thoroughly in a mixture of 0.05% Tween-20 in PBS. The sections were further incubated with 12-nm colloidal gold-conjugated goat anti-rabbit (1:50; Jackson ImmunoResearch, West Grove, PA) goat anti-mouse IgG (1:30; Jackson ImmunoResearch), or 6-nm colloidal gold-conjugated donkey anti-sheep IgG (1:100; Jackson ImmunoResearch) at RT for 1 hour. The sections were then rinsed, stained with uranyl acetate, and examined under a transmission electron microscope (model JEM-1220; JEOL, Peabody, MA) at 80 kV accelerating voltage. As a negative control, normal rabbit, mouse, or sheep IgG or anti-myocilin preadsorbed with the immunogenic peptide was used instead of the primary antibody at the equivalent concentration of IgG fraction.
To localize hyaluronic acid, the sections were incubated with biotinylated hyaluronic acid binding protein (HABP; 1:10; United States Biological) after blocking. The sections were next incubated for 1 hour in horseradish peroxidase (HRP)-conjugated streptavidin (1:50; Jackson ImmunoResearch) and for another hour in 6-nm colloidal gold-conjugated goat anti-HRP (1:20; Jackson ImmunoResearch).
To determine the relative distribution of each of the 16 molecules, the intensity of immunogold labeling or the amount of gold particles in each ECM structure was graded. More than 10 micrographs from each eye were examined by two masked observers. The intensity was scored from ± to +++, with ± representing minimal staining and +++, intense staining.
To examine the interaction between myocilin and other ECM components, double labeling was performed. One side of the section was immunostained for myocilin and labeled with 12-nm gold particles, and the other side was stained for ECM elements with 6-nm gold particles. Single-side incubation was achieved by floating the grids on a drop of each solution.