Biological Properties of Trabecular Meshwork Cells

Abstract

The molecular and physiological mechanisms that lead to the progression of glaucoma are poorly understood. Despite the fact that glaucoma afflicts millions of people worldwide, research on the disease is limited by the current animal models that do not translate well to human forms of the disease. However, recent advances in culturing and manipulating human trabecular meshwork cells may provide a means to elucidate some of the mechanisms that cause glaucoma. This review focuses on the properties of trabecular meshwork cells, from their characteristic expression profile in vivo to their responsiveness to biochemical and biophysical signals in vitro. Hopefully the study of cultured trabecular meshwork cells will provide a better understanding of glaucoma and lead to new, much needed therapies.

Introduction

The trabecular meshwork (TM) of the eye, composed of cells and matrix, is thought to regulate aqueous humor outflow to control intraocular pressure (IOP). The TM and aqueous humor outflow pathway have a wide anatomical variation from one species to another(Chen et al., 2008; Samuelson, Gelatt, 1984; Smelser, Ozanics, 1971); however, the trabecular meshwork of human, rat and mouse all contain extracellular meshwork organized into a network of beams covered with endothelial cells. There are extensive numbers of elastic-like fibers in the juxtacanalicular (JCT) or cribiform region of the human trabecular meshwork (HTM) that form connections to the inner wall of Schlemmm's canal. These fibers appear to function as mechanical tethers to maintain contact of the JCT to the inner wall, resulting in the opposition of hydrodynamic forces generated during perfusion. For this reason, the human meshwork appears unique in that it does not have the perfusion-volume-dependent increase in aqueous outflow facility known as “washout”(Scott et al., 2007). Species variations in the meshwork make the study of glaucoma difficult since no good animal model for human disease exists.

Due to the differences in species, this review article will primarily focus on the human trabecular meshwork. However, there is still complexity and variation to consider within the same individual. Tshumper and Johnson found notable differences between fellow eyes(Tschumper, Johnson, 1990) in terms of absolute cell numbers as well as cellularity (nuclei of HTM cells/solid tissue area). The average variation in cell numbers was 15% with deviations up to a maximum of 37%. Interestingly, data also indicate that the cellularity difference may be higher in patients with primary open angle glaucoma (Alvarado et al., 1984). Despite these variables, studying the properties of HTM cells remains a valuable tool to better understand glaucoma.

Culture Systems

Due to the lack of model systems to study glaucoma, the use of cultured TM cells provides a means to evaluate their biological properties when challenged with conditions linked to glaucoma(Polansky et al., 1979; Polansky et al., 1984; Tripathi, Tripathi, 1982; Weinreb et al., 1983). HTM cells are usually obtained from whole donor eyes or from corneal buttons and ideally, one would like to compare HTM cells from glaucomatous tissue to normal HTM cells. (Rhee et al., 2003) However, the culturing of cells from donor glaucomatous tissue has presented a challenge because of limited cellular replication in vitro. Certain groups have had some success using different methods to isolate the HTM cells from glaucomatous meshwork(Stamer et al., 2000; Wordinger et al., 1998), but in general, culturing of HTM cells from glaucomatous meshwork is much more demanding than obtaining and maintaining cells from normal donors. Therefore, normal HTM cells are often stimulated with different techniques to induce glaucoma-like phenotypes in vitro. These techniques will be reviewed later.

HTM Cell Characteristics

Perhaps the most characteristic marker for HTM cells is the increased expression of myocilin after the cells have been treated with dexamethasone(Polansky et al., 2000b). The induction appears to be specific in HTM cells compared to other ocular cells. This protein has characteristics of a matricellular protein(Peters et al., 2005; Shen et al., 2008; Wentz-Hunter et al., 2004). Other proteins that have been identified in HTM cells are: the alpha-2 adrenergic receptor(Stamer et al., 1996), aquaporin-1(Stamer et al., 1995) and acetylated and acetoacetylated low-density lipoproteins(Chang et al., 1991), but these proteins are also present in a wide variety of cells and are not specific markers for HTM cells. One protein, αB-crystallin, has been reported to be present in cells of the JCT, but not in other cells of the HTM(Welge-Lussen et al., 1999). This heat shock protein can be induced with transforming growth factor-β(TGF-β) as well as heat shock and oxidative stress(Tamm et al., 1996). Because this protein is expressed solely in the JCT region, there is a possibility that there are different sub-types of cells within the meshwork. However, no data exist at this time that differentiates distinct cell types within the HTM. Although it is difficult to use a single marker protein to identify a HTM cell, they do have a pronounced phagocytosis rate that can be used as a behavioral characteristic.

HTM Cell Phagocytosis

The elevated intraocular pressure that often accompanies glucocorticoid treatment has been linked to increased deposition of extracellular matrix material in the outflow pathway(Wordinger, Clark, 1999). This deposition is in part due to decreased HTM cell phagocytosis after steroid treatment(Matsumoto, Johnson, 1997; Zhang et al., 2007). These data highlight phagocytosis as an important function for HTM cells. A variety of particles such as latex beads, zymosan particles, or Staphylococcus aureus bioparticles have been used in animal studies in vivo, perfused organ culture of donor human eyes, and HTM cells in vitro.(Buller et al., 1990; Johnson et al., 1989; Sherwood, Richardson, 1988; Zhang et al., 2007) (Figure 1). In a study using cats, one eye received phagocytic challenge in organ culture while the fellow eye was challenged in vivo(Buller et al., 1990). There were more cells involved with phagocytosis in vivo suggesting an activation of cells possibly by an inflammatory pathway that could only be possible in vivo. There also appeared to be a loss of fibronectin and laminin, indicative of phagocytosis mediated ECM remodeling (Zhou et al., 1995).

Figure 1.

Primary HTM cells from non-glaucomatous donors show high rates of phagocytosis in culture. HTM cells were plated and challenged with particles that emit fluorescence (green) when actively internalized. Four hours after challenge, a high percentage of the particles were phagocytosed by the HTM cells.

Gamma-interferon has also been shown to block phagocytosis(Park, Latina, 1993). This inhibition is apparently related to the alterations of actin filaments in the cell that form radial spoke-like arrangements. Conversely, platelet-derived growth factor enhanced phagocytosis by limiting stress fiber formation in TM cells(Tamura et al., 1989). Altogether, these data provide a correlation between the regulation of intraocular pressure to phagocytosis and the remodeling of the extracellular matrix in HTM cells.

Two of the common ways to study the HTM and progression of glaucoma are tissue culture of HTM cells and the TM organ culture of human, primate and other species within the anterior segment of the eye(Bahler et al., 2008; Comes, Borras, 2007; Johnson, Tschumper, 1987; Keller et al., 2008; Pang et al., 2000; Wang et al., 2008). In general, hypothesis testing related to glaucoma is first explored with cultured cells and then confirmation of the data is performed with the more difficult organ culture system. This review focuses on the former.

Cell-based Glaucoma Models

HTM cells grown in vitro do not always behave similar to their in vivo counterparts. The changes in gene expression and protein production in tissue culture systems can make it particularly challenging to use HTM cells for glaucoma models. As a result, four different models employing external stimuli have been developed that allow researchers to use normal HTM cells to study glaucoma progression using an in vitro environment. Even though these models recreate some glaucoma-like properties in culture, it is important to note that all of these models still have limitations.

Dexamethasone

One of the earliest HTM-based glaucoma models, and perhaps the most widely studied to date, involves the use of glucocorticoids. Glucocorticoids are a family of signaling molecules that produce a wide range of effects depending on the receptors they bind and transduction cascades they induce. Glucocorticoids, like dexamethasone, can increase intraocular pressure, and this finding was then translated to cell culture to model glaucoma-like changes in HTM cells(Polansky et al., 2000a).

Dexamethasone has been shown to cause a decrease in the amount of endothelin-1 receptor B (ET-1rB) present in HTM cells grown in culture(Zhang et al., 2003). If ET-1rB is allowed to bind endothelin-1 under normal conditions, it drives nitric oxide release and subsequent vasodilation(Hirata et al., 1993). Thus, dexamethasone treatment decreases the level of ET-1rB available resulting in a drop of nitric oxide levels. Without ET-1rB present, only endothelin-1 receptor A remains available for the ligand. The binding of endothelin-1 to receptor A can start a vasoconstriction signaling cascade, which would presumably promote less outflow in the TM and raise IOP.

Dexamethasone has been reported to induce other hallmarks of glaucoma in HTM cells. Glucocorticoids cause the formation of cross-linked actin networks in cultured HTM cells(Clark et al., 1994). More rigid HTM cells with dense actin networks are less pliable and more likely to inhibit outflow. Dexamethasone downregulates matrix-metalloproteases (MMP)-2, -3, and -14 when the cells are subjected to high pressures (50 mm Hg)(Ehrich et al., 2005). This results less ECM turnover and less outflow. Untreated HTM cells subjected to 50 mm Hg pressure had a strong upregulation of MMP-3(Ehrich et al., 2005). It appears the normal response of HTM cells to elevated pressure is to increase the amount of MMP-3 to help degrade and remodel ECM, but the presence of dexamethasone blocks this effect. TGF-β

The second glaucoma model for HTM cells is dependent on treatment with TGF-β. TGF-β is a super-family of signaling proteins that have also been found at high levels in the aqueous humor of glaucomatous eyes(Tripathi et al., 1994). It is believed that high levels of TGF-β2 contribute to the progression of POAG. TGF-β can exist in active and latent forms, but it has been shown that thrombospondin-1, a TGF-β activator, is upregulated in glaucomatous tissue and may initiate the signaling cascade(Flugel-Koch et al., 2004). In cultured HTM cells, 1 ng/ml of either TGF-β1 or β2 will increase tissue transglutaminase, which can cross-link ECM proteins(Welge-Lussen et al., 2000). TGF-β will also promote fibronectin production. Both outcomes can modify the ECM and decrease outflow. However, some of the bone morphogenic proteins (BMPs), which are part of the TGF-β super family, can balance TGF-β effects in the TM. BMPs and BMP receptors have been found in the TM and TM cells(Wordinger et al., 2002). There are some specific differences, though, notably BMP-2 was present at a much higher level in native tissue than in cell culture(Wordinger et al., 2002). While TGF-β induced fibronectin expression in HTM cells, co-treatment with BMP4 inhibited fibronectin expression. TGF-β2 induced collagen IV and VI, but that expression was blocked by co-treatment with BMP-7(Fuchshofer et al., 2007). The delicate balance of TGF-β and BMP proteins may play a large role in properties of the ECM and ultimately, the pathogenesis of glaucoma.

Oxidative Stress

The third model employs oxidative stress to mediate a glaucoma-like phenotype in HTM cells. Chronic oxidative stress can contribute to reduced outflow by inhibiting the intracellular proteasome system that works to degrade intra- and intercellular debris. After exposure to oxidative stress (40% oxygen) for 10 days, HTM cells lose proteasome activity and have increased cell death(Caballero et al., 2003). Much like the balance between TGF-β and BMPs, there is a balance within the oxidative stress system. Continuous exposure of HTM cells to oxidative stress via H₂0₂ results in ROS generation. This, in turn, stimulated NF-κB activation and subsequent production of interleukins (ILs), namely IL-1(Li et al., 2007). The positive feedback loop between NF-κB activation and IL-1 expression may be a response to lower IOP; however, when IL-1 was administered exogenously, aqueous outflow in rat eyes decreased(Kee, Seo, 1997). Interestingly, normal HTM cells had activated NF-κB and increased levels of IL-1 after treatment with phacoemulsification and this procedure often lowers IOP. This supports the hypothesis that this inflammatory pathway is a cellular attempt to decrease IOP(Wang et al., 2001). The complex and broad influence of proteasome stimulation and inflammation response means more research needs to be performed to understand the role oxidative stress may play in glaucoma.

Mechanical Stretch

Elevated IOP exerts physical forces on HTM cells causing mechanical stretch. As the fourth model for glaucoma, stretching HTM cells can have a profound effect on their expression profile. A gene array study that examined 8,000 possible genes found that mechanical stretch applied over 12, 24, or 48 hours results in a downregulation of 29 and an upregulation of over 100 genes in HTM cells(Vittal et al., 2005). Tenascin C, fibronectin, TGF-β receptor complex protein, secreted protein, acidic, cysteine-rich (SPARC) and some collagens were upregulated 1.5 - 3.7 fold in the array. Cytoskeleton proteins vimentin, alpha 1 tubulin, and SH2-B as well as inflammatory protein NF-κB (p105) were also upregulated after stretch. There are also reports that mechanical stretch increases the expression of MMP-2 and MMP-14 through the intracellular signaling molecule, mTOR(Bradley et al., 2001; Bradley et al., 2003). The MMP upregulation, in turn, modulates remodeling of the ECM.

Besides controlling transcription levels of many genes, mechanical stretch can mediate the alternative splicing of several mRNAs as well. Tenascin C, collagen XII and CD44 are alternatively spliced in response to stretch and change the trabecular meshwork ECM composition(Keller et al., 2007). Stretch exerts great control over fibronectin splicing(Vittal et al., 2005). Different fibronectin domains reversibly alter outflow in TM cultures through the available binding domains in the molecule(Santas et al., 2003).

Biophysical Cues

While biochemical signals are commonly used in laboratory experiments to control cell behaviors, the mechanical stretch model demonstrates the importance of also studying the effects of biophysical signals. In addition to fluctuations in IOP, HTM cells are provided with thousands of biophysical cues through cell-cell interactions and cell-ECM interactions. ECM has a rich, 3-dimensional topography through which the cells grow and function(Brody et al., 2006; Diehl et al., 2005; Liliensiek et al., 2006). HTM cells in vivo never see the flat surfaces on which they are typically cultured in vitro. When HTM cells were seeded onto nano and micro scale substrates that biomimetically share topographic features with native ECMs and basement membranes, cellular behaviors and gene expression were altered(Russell et al., 2008). The actin cytoskeleton, and overall HTM cell orientation aligned with the artificial substrata provided in the form of repeating nano and microscale grooves and ridges. In addition, HTM cells dramatically elongated parallel to the anisotropic biomimetic substrates (Figure 2). Finally, myocilin expression was upregulated on topography independent of dexamethasone stimulation(Russell et al., 2008).

Figure 2.

HTM cell alignment is modulated by underlying topographic cues. Cells were cultured for 24 hours prior to fixation, phalloidin staining and imaging. Cells oriented between 0-10° to the repeating ridge and groove pattern of the biomimetic substrate were defined as parallel. On the control planar surface, the cells were equally likely to be aligned parallel as perpendicular to an arbitrary, consistent axis.

The topography of the underlying substrata is not the only physical cue that HTM cells sense. The compliance (rigidity) of the substrate can also modulate HTM cell behaviors and interactions. HTM cells seeded on polyacrylamide gels spread less distance and at a slower rate than HTM cells plated onto stiff tissue culture plastic(Schlunck et al., 2008). The actin cytoskeleton formed less pronounced stress fibers and focal adhesion kinase is minimally phosphorylated on the soft gels compared to plastic. Similar to the topographic surfaces, myocilin expression was upregulated on the more compliant gels. These data, combined with the experiments using biomimetically topographic surfaces reveal the importance of incorporating biophysical cues into HTM cell culture studies.

Summary

The properties of HTM cells are complex in nature and still require additional research. While they provide a useful starting point to understand the different facets of facility and pathogenesis of glaucoma, we must realize that models still have limitations. Aspects of the different HTM in vitro models can be combined and developed to create more accurate representations of TM cell behavior in vivo. The incorporation of biophysical cues will also play an important role in the development of our research systems. Ultimately, a more complete grasp of the cellular properties of HTM cells will provide us with new treatments for glaucoma.