Expression of Wild-Type and Truncated Myocilins in Trabecular Meshwork Cells: Their Subcellular Localizations and Cytotoxicities

Abstract

purpose. To investigate the subcellular localizations and potential cytotoxicities of wild-type and truncated (Q368X) myocilin in cultured human trabecular meshwork (TM) cells.

methods. Full-length wild-type myocilin, truncated myocilin, and stromelysin were expressed as green fluorescence (GFP) or DsRed fusion proteins in TM cells by using adenoviral vectors, and their secretory properties and cytopathic effects were evaluated by Western blot analysis and cell proliferation assay, respectively. To determine the subcellular localizations of myocilins, the cellular organelles of the infected TM cells were stained with organelle-specific antibodies or fluorescent indicators and examined under a confocal microscope.

results. Wild-type myocilin was expressed mainly in the perinuclear region of TM cells and was localized preferentially in endoplasmic reticulum (ER), but not in actin, microtubules, or mitochondria. Truncated myocilin was also localized in ER, and its expression was found to be potentially toxic to TM cells, leading to deformed cellular morphology and diminished cell proliferation, but it had no effect on the secretion of stromelysin. The truncated myocilin was also found to be colocalized with and appeared to aggregate with wild-type myocilin when the proteins were coexpressed.

conclusions. TM cells participating in the turnover of trabecular extracellular matrix (ECM) components are important in regulating aqueous outflow. The truncated myocilin, colocalized and coaggregated with wild-type myocilin, is believed to cause a dysfunction of the cells, resulting in alterations in structural compartmentalization of trabecular ECM and obstruction of aqueous outflow.

Glaucoma, a complex ocular disease characterized by progressive optic neuropathy, is the leading cause of irreversible blindness worldwide. Primary open-angle glaucoma (POAG) is the most common form and is generally associated with elevated intraocular pressure (IOP) consequent to the abnormal resistance of aqueous outflow through trabecular meshwork (TM), a specialized tissue lining the outflow pathway of the eye. ¹  

Myocilin is a glycoprotein with a molecular weight of approximately 55 kDa and is composed of two major domains, an N-terminal myosin-like domain (thus named myocilin) and a C-terminal olfactomedin-like domain. ^{2 3 4} The protein, formerly referred to as trabecular meshwork–inducible glucocorticoid protein (TIGR), is of particular interest in glaucoma genetics because of a suggested link between the protein and POAG. The evidence for the link includes (1) the induction profile of myocilin by glucocorticoid (GC), which mimics GC-induced glaucoma in dose response and time course, ^{5 6 7} (2) localization of the gene encoding the protein on 1q23-q25, ⁸ a GLC1A locus identified by genetic linkage analysis, ^{9 10 11 12} and (3) occurrence of various mutations associated with POAG in the gene. ^{8 13 14 15 16} It has been assumed that myocilin is one of the aqueous outflow regulators in outflow cells, because the protein is normally present in both the trabecular cells of the uveal, corneoscleral, and juxtacanalicular regions and the endothelial cells of Schlemm’s canal. ^{17 18 19} Furthermore, the perfusion of recombinant myocilin into human eye organ culture increases outflow resistance. ²⁰ However, the biological activities and physiological functions of myocilin remain largely unknown. 

One of the distinguishing characteristics of myocilin is that, whereas the normal form is completely Triton soluble, mutant forms are substantially insoluble. ²¹ Consistently, there have been reports that normal myocilin is well secreted extracellularly, but little to no myocilin is secreted from cells that express different mutant forms of myocilin. ²² Moreover, it has been recognized that the expression of truncated myocilin results in reduced secretion of normal (endogenous) myocilin. ^{22 23} A recent study has revealed that the characteristics of truncated myocilin may be related to its incorrect processing in the endoplasmic reticulum (ER) and accumulation into insoluble aggregates, suggesting that the mutant myocilin acts intracellularly in the development of POAG. ²⁴ To better understand the role(s) played by myocilin in glaucoma, it is necessary to determine the exact subcellular localization of myocilin in TM cells, but published results on this question have not been in agreement with each other. ^{25 26 27 28}  

In the present study, we reexamined the subcellular localization of virally expressed normal and truncated (Q368X) myocilins in TM cells. During that work, we noted that the truncated myocilin showed a potential cytopathic effect in the cells. The present study was therefore undertaken to investigate the subcellular localizations and cytotoxicities of both wild-type and truncated myocilins in cultured human TM cells. 

Materials and Methods

Cell Culture

All cells used in this study were cultured in Dulbecco’s modified Eagle’s medium (DMEM, Gibco, Grand Island, NY) supplemented with 10% fetal bovine serum (FBS; Gibco) and antibiotics. TM cells were a generous gift from Jon Polansky (University of California, San Francisco, CA). 

Adenoviral Vector Constructions

The cDNAs for truncated and full-length wild-type myocilin were amplified separately by reverse transcription–polymerase chain reaction (RT-PCR), using total RNA extracted from TM cells treated with 250 nM dexamethasone for 7 days. The sequences of primers used were 5′-AGGAAGCCTCACCAAGCCTC (sense primer, nucleotides [nt] 13–32 in GenBank accession number U85257 for nt position numbering; GenBank is provided in the public domain by the National Center for Biotechnology Information, Bethesda, MD, and is available at http://www.ncbi.nlm.nih.gov/Genbank), 5′-ACCGGTCCGTGGTAGCCAGCTCCA (antisense primer for truncated myocilin, nt 1138–1119), and 5′-ACCGGTAACATCTTGGAGAGCTTGA (antisense primer for wild-type myocilin, nt 1551–1532). The antisense primers contain four or five extra mucleotides in its 5′ end to create a flanked Age1 site and a change from C to A at nt 1550 to destroy the stop codon. The amplified cDNAs were verified by sequencing and cloned into the EcoRI and Age1 sites of the shuttle vector pΔACMV-GFP, ²⁹ resulting in the pΔACMV-truncated myocilin-GFP and pΔACMV-wild-type myocilin-GFP vectors, respectively. The green fluorescence protein (GFP) gene was originated from the pEGFP-N1 vector (Clontech, Palo Alto, CA) digested with EcoRI and Not1. 

The vector expressing the myocilin with red fluorescence was generated by replacing the GFP gene in the pΔACMV-wild-type myocilin-GFP vector with the DsRed1 gene obtained by digestion of the pDsRed1-N1 vector (Clontech) with Age1 and Not1. 

The cDNA for human stromelysin-1 was amplified by PCR, as previously described, ²⁹ except that the antisense primer 5′-ACCGGTAAACAATTAAGCCAGCTGT (nt 1477–1458 GenBank accession number X05232) was used. The primer also contains five extra nucleotides in its 5′ end to create an Age1 site and a change from C to A in nt 1476 to destroy the stop codon. The cDNA was verified by sequencing and cloned into the pΔACMV-GFP vector. 

Replication-deficient recombinant adenoviruses were generated as described previously. ²⁹ Briefly, the constructed plasmids described in the prior paragraph were separately transfected into 293 cells, along with an adenovirus-5 genome-containing plasmid, pJM17, using a lipid-based transfection reagent (Lipofectamine; Gibco). Viral particles were obtained after 7 to 9 days and propagated in 293 cells. The viruses with the desired gene construction were identified by PCR, and viral titer was determined by limiting dilution on 293 cells. Contamination of the replication-competent virus was ascertained by infecting Cos-7 and HeLa cells with each viral stock. There was no sign of replication of the viruses in the cells. 

Western Blot Analysis

TM cells of passages before passage 8 were seeded in 12-well plates and cultured with 0.5 mL medium to reach approximately 70% confluence. The cells were infected with the viruses at a multiplicity of infection (MOI) of 50 to 100 plaque forming units (pfu) per cell for 2 hours, washed once with culture medium, replenished with DMEM containing 2% FBS, and further cultured. After 48 hours, the culture medium was harvested and centrifuged at 17,000 rpm for 10 minutes, and the cleared supernatants were mixed with 4× sodium dodecyl sulfate (SDS) sample buffer. The plate with the adherent cells was washed two times with PBS and an equivalent amount of 1× SDS sample buffer was added (0.6 mL). The cell lysates were harvested and cleared by centrifugation. 

SDS polyacrylamide gel electrophoresis was performed according to the method of Laemmli. ³⁰ The samples were boiled for 5 minutes, and 10 μL of each sample (5 μg) was loaded onto 10% polyacrylamide gels and electrophoresed at 100 V for 1 hour. The resolved proteins were transferred electrically to a nitrocellulose membrane, blocked overnight with 0.5% skim milk in Tris-buffered saline containing 0.05% Tween 20 (TTBS). The membrane was incubated for 2 hours with polyclonal rabbit anti-GFP antibodies (Clontech) diluted 1:2000 in TTBS and washed three times with TTBS for 10 minutes between washes. The membrane was probed with peroxidase-conjugated anti-rabbit antibody (Amersham, Piscataway, NJ) diluted 1:5000 in TTBS for 2 hours and developed with a chemiluminescence detection kit (Amersham, Amersham UK), according to the manufacturer’s guide. 

Organelle Staining and Confocal Microscopy

TM cells were seeded onto four-well chamber slides at a confluence of 30% and infected with viruses at an MOI of 500 pfu for 2 hours. The cells were washed briefly, and further cultured for 48 hours. Live cell staining for Golgi apparatus, mitochondria, and lysosomes was performed with vital stains specific for each organelle (Bodipy TR ceramide D-7540, 5 μM; MitoTracker M-7512, 300 nM; and LysoTracker L-7528, 300 nM, all purchased from Molecular Probes, Eugene, OR) The stains were added to the cultures and incubated for 30 to 60 minutes according to the manufacturer’s procedure. Staining for actin microfilaments was performed by adding a lipophilic fluorophore (250 nM; Bodipy, B-7464; Molecular Probes) to cells that had been fixed with 4% paraformaldehyde in phosphate-buffered saline (PBS) for 10 minutes and permeabilized with 0.5% Triton X-100 in PBS for 5 minutes. To stain microtubules and the ER, the cells were fixed and permeabilized, and incubated for 1 hour with monoclonal anti-tubulin antibody (Sigma, St. Louis, MO) diluted 1:250 in PBS containing 2% bovine serum albumin or anti-protein disulfide isomerase (PDI; Stressgen, Victoria, British Columbia, Canada) diluted 1:50. After they were washed briefly with PBS, the cells were incubated for 30 minutes with tetramethylrhodamine isothiocyanate (TRITC)–conjugated secondary antibody (Zymed, South San Francisco, CA) diluted 1:50 in PBS. 

Images of green myocilin fluorescence were collected with a confocal laser scanning microscope (LSM 510; Carl Zeiss, Jena, Germany) using 488-nm excitation light from an argon laser and a 505- to 530-nm bandpass barrier filter. Images of red actin, tubulin, ER, Golgi, mitochondria, and lysosome fluorescence were collected using a 543-nm excitation light from the HeNe laser, a 560-nm dichroic mirror, and a 570-nm long-pass filter. Cells were illuminated only during image acquisition (3.7 seconds/frame for GFP and for TRITC) and the images were collected as a Z-series of 5 to 10 serial images at 0.5-μm intervals. 

Cell Proliferation Assay

TM cells plated in a 96-well plate at a concentration of 1 × 10⁴ cells per well were infected with viruses at an MOI of 15 to 1000 pfu for 2 hours. The cells were washed, replenished with 100 μL culture medium, and further cultured. After 72 hours, 10 μL of cell proliferation reagent (WST-1; Roche, Mannheim, Germany) per well was added, and absorbance at 450 nm was measured after 1 hour. 

Results

Characteristics of Myocilins Tagged with GFP

Throughout this study, adenoviral vectors were used to express the desired proteins in TM cells that were robustly transfected by naked DNA. Five kinds of replication-deficient recombinant adenoviruses were generated, each carrying wild-type and truncated myocilin, stromelysin, or GFP. The first three proteins were designed expressed as GFP or DsRed fusion proteins by placing the cDNA encoding each protein to the N terminus of the fluorescent proteins to facilitate direct visualization (Fig. 1) . 

By Western blot analysis, it was confirmed that the tagging of proteins with a relatively large GFP did not affect the secretory property of each native protein. As shown in Figure 2 , truncated myocilin-GFP was resolved at approximately 68 kDa and was detected only in cell lysates, whereas wild-type myocilin-GFP of approximately 83 kDa was detected in both cell lysates and culture medium when expressed solely, but only in cell lysates when coexpressed with truncated myocilin-GFP. Similarly most stromelysin-GFP was secreted into the culture medium. These results coincided well with findings previously reported. ^{22 31} It should be noted, however, that the secretion of stromelysin was not significantly reduced when coexpressed with truncated myocilin. 

Subcellular Localization of Myocilins

In TM cells, wild-type myocilin was expressed as discrete fine vesicular structures scattered mainly in the perinuclear regions, coinciding with previous observations. ^{25 28} However, the subcellular localization of the myocilin was not consistent with any of the previous reports. ^{25 26 27 28} As seen in Figure 3 , the protein was localized preferentially in the ER and partially in the Golgi apparatus, but gave no evidence of localization to actin filaments or mitochondria. There was also no clear colocalization of myocilin with microtubules, although sparse overlapping between GFP fluorescence and tubulin fiber staining was observed, especially in the microtubule-organizing center near the nucleus. Myocilin did not show colocalization with lysosomes. 

Truncated myocilin was also localized to the ER (Fig. 4) , but not to the other cytoplasmic organelles examined (data not shown). The protein, however, appeared to show more aggregation than did its wild-type form and induced PDI, an ER-resident enzyme, to aggregate. The aggregated PDI which otherwise was dispersed evenly in the ER implied that the enzyme was bound to and trapped within the myocilin aggregates. 

The truncated myocilin was also found to induce wild-type myocilin (tagged with DsRed) to aggregate (Fig. 5) . The wild-type and truncated myocilins may interact with each other by the formation of a dimmer or oligomer. The interactions appeared to be the aggregation of truncated and wild-type myocilin and thus resulted in the colocalization of both proteins. 

Cytotoxicity of Truncated Myocilin

The visualization of GFP fusion protein under a fluorescence microscope necessitates a viral infection at an MOI higher than 100 pfu. In this condition, it was found that the expression of truncated myocilin did not show dose responsiveness and seemed to be lower than expression of other proteins (data not shown). This phenomenon, also noted by other investigators, ^{22 32} was thought to be attributable to the potential cytotoxicity of the truncated myocilin, in that the shape of cells expressing the protein was severely changed, as was clearly revealed by light microscope (Fig. 6) . To confirm this possibility further, a cell proliferation assay was performed on the TM cells infected with each adenoviral vector at an MOI of up to 1000 pfu. Figure 7 clearly shows that truncated myocilin started to exert its cytopathic effect even at an MOI of less than 100 pfu, compared with other proteins. By contrast, wild-type myocilin showed cytotoxicity comparable with GFP expression that was also slightly toxic to the cells. As expected, stromelysin, a well-known secretory protein, ³¹ was the least toxic to the cells. 

Discussion

More than 40 different myocilin mutations have been described in the literature. The Q368X mutation occurs most commonly in patients with POAG, ^{8 13 15} but little is known about the mechanisms by which the mutation causes POAG. In the current study, truncated myocilin exhibited potential cytotoxicity in trabecular cells, giving clues to understanding how the myocilin is involved in the pathogenesis of POAG. It is well established that outflow resistance increases with age, accompanied by a decreased number of TM cells and alterations of the juxtacanalicular extracellular matrix (ECM). ^{33 34 35} It is also well known that glaucomatous eyes exhibit fewer TM cells and an abnormal appearance of ECM in the juxtacanalicular region compared with age-matched control eyes. ^{36 37} Therefore, the cytotoxicity of the truncated myocilin that inevitably results in diminished number and dysfunction of trabecular cells of the outflow tissue, may lead to decreased outflow facility and ultimately to the development of POAG. 

The reason for the cytotoxicity of truncated myocilin was not intensely investigated in the current study, but it can be postulated that the aggregated expression of the myocilin may be one of the causes. It is highly probable that some mutations in myocilin cause the protein to be folded and processed inadequately. If this is true, the mutant myocilins not only lead to loss of normal functions but also undergo proteolytic degradation. ^{38 39} The subsequent accumulation of the degraded products may support the implication that they are cytotoxic and contribute to degenerative disease marked by the formation of insoluble protein plaques. Alzheimer’s disease, characterized by presence of plaques and tangles in the brain, is a good example. ⁴⁰ There has been biochemical evidence concerning misprocessing of truncated myocilin, including its inefficient processing and accumulation in the ER, ²⁴ poor expression in certain cell types (e.g., A549 cells), ²² or expression only after treatment with lactacystin, ³² a proteasome inhibitor. The novel finding in this study of coaggregation of PDI with the truncated myocilin provides additional evidence, because the enzyme occasionally forms a disulfide bond with the irreversibly misfolded proteins. ⁴¹ It is also tempting to speculate whether E323K mutation that causes a translocational pause near the beginning of the olfactomedin domain generates myocilin products mimicking the truncated form with altered processing. ^{42 43}  

The colocalization of truncated myocilin with wild-type myocilin was not surprising, in that myocilin can exist as either a dimer or oligomer, ^{4 44} and it is possible that oligomerization occurs between wild-type and truncated myocilins. ⁴⁵ The colocalization demonstrates a place where the two proteins may interact. In addition, the coaggregation of wild-type myocilin due to the interaction explains why the secretion of wild-type myocilin is suppressed when coexpressed with truncated myocilin, whereas the normal secretion of stromelysin demonstrates that the central secretory pathway was not disrupted by the aggregation of the truncated myocilin. 

Previous studies have localized wild-type myocilin in TM cells to microtubules, ²⁵ mitochondria, ^{26 28} or cross-linked actin networks. ²⁷ However, no localization of wild-type myocilin to any of these organelles was found on the current study. At present, the reason for the differences between our results and those of others has not been clarified, but it may be related to the variations in cell types examined or the detailed experimental procedures used. For example, in our experience the red fluorescent paclitaxel used in the study by Mertts et al. ²⁵ always overstained the tubulin fiber. The other factor for special consideration is the expression level of the myocilin. Sometimes, subcellular localization of certain proteins requires a high level of protein expression, and it is possible that the overexpressed proteins undergo modified processing and different localizations, as discussed by Ueda et al. ²⁶ Nevertheless, our result showing localization of myocilin to the ER is in good agreement with the secretory property of myocilin. It was unexpected, however, that the myocilin showed only partial overlapping to the Golgi apparatus. Further investigation may be necessary to investigate this question. 

Of the true role(s) played by wild-type myocilin in outflow cells and tissues, one possibility is that overexpression of myocilin is a cause of POAG. ^{6 7} Supporting this hypothesis are the findings of GC inducibility of the myocilin, ^{4 5} a greater degree of myocilin staining in eyes with POAG than in age-matched control eyes, ⁴⁶ and increased outflow resistance in the anterior segment organ culture by infusion of recombinant wild-type myocilin. ²⁰ In our study, truncated myocilin exhibited cytotoxicity in the trabecular cells, an effect postulated to result in the diminished number and dysfunction of the trabecular cells, thereby decreasing the outflow facility and contributing to development of glaucoma. It is currently not known whether other frequently occurring mutant myocilins besides truncated myocilin suppress the secretion of wild-type myocilin and exhibit cytotoxicity. 

Although other investigators believe that secretion of wild-type myocilin is needed for the maintenance of normal aqueous outflow, there was little evidence to support this in our study. This hypothesis may be explained by the finding that truncated myocilin inhibits secretion of wild-type myocilin. ^{22 23} More research must be concentrated on the contradictions in our present understanding to better elucidate the mechanism of POAG. This is necessary, because most patients with POAG do not carry a myocilin mutation; nonetheless, they have glaucoma. A complete appreciation of the role played by myocilin in outflow regulation will greatly contribute to the understanding of glaucoma. 

 Supported by a grant from the Korea Science and Engineering Foundation, SRC Project endowed to the MTRC.

 Submitted for publication May 13, 2002; accepted June 28, 2002.

 Commercial relationships policy: N.

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Schematic diagrams of wild-type myocilin (wt MYOC), truncated myocilin (mt MYOC), and stromelysin constructs with amino acid numbers depicted. Pcmv, cytomegalovirus promoter; pA, simian virus (SV)40 polyadenylation signal.

Figure 2.

 

Western blot analysis of adenovirally expressed proteins in TM cells. (A) Truncated myocilin-GFP (mt MYOC), wild-type myocilin-GFP (wt MYOC), and stromelysin-GFP (stromelysin) were expressed in cultured human TM cells. Viral infection was performed for 2 hours at an MOI of 50 (left lanes) and 100 (right lanes) pfu. The cells and culture medium were separately harvested 48 hours after the infection, and an equivalent amount of each sample was resolved on an acrylamide gel, transferred to a nitrocellulose membrane, and probed with anti-GFP antibody. (B) TM cells were infected with the viral vectors encoding wild-type myocilin-GFP or stromelysin-GFP at an MOI of 50 pfu, followed by infection with the vector encoding truncated myocilin-GFP at an MOI of 25 (left lane) and 50 (right lane) pfu. The influence of truncated myocilin on the secretion of stromelysin was examined.

Figure 3.

 

Subcellular localization of wild-type myocilin in TM cells. TM cells were infected for 2 hours with an adenoviral vector encoding wild-type myocilin-GFP at an MOI of 500 pfu. After 48 hours, each cytoplasmic organelle of the cells was stained and examined under a confocal microscope. Each row represents the same cell. The staining patterns of each organelle (red) and myocilin-GFP (green) are shown in the left and middle columns, respectively. The overlays of the respective pictures are shown in the right column. The colocalization generates a yellow stain. Scale bar, 10 μm.

Figure 4.

 

Coaggregation of PDI with truncated myocilin in TM cells. The image of PDI staining (A) and truncated myocilin-GFP (B) were merged (C). The experimental procedures were the same as in Figure 3 . Scale bar, 10 μm.

Figure 5.

 

Colocalization and coaggregation of wild-type myocilin with truncated myocilin in TM cells. Wild-type myocilin-DsRed (red) showed punctate, fine, vesicular structures when expressed solely in TM cells (A). However, the wild-type myocilin became aggregated (B) when coexpressed with truncated myocilin-GFP (C) and showed colocalization with the truncated myocilin (D). Scale bar, 10 μm.

Figure 6.

 

Cytotoxicity of truncated myocilin in TM cells. TM cells were infected for 2 hours with each adenoviral vector carrying truncated myocilin-GFP (mt MYOC), wild-type myocilin-GFP (wt MYOC), stromelysin-GFP (stromelysin), or GFP at an MOI of 500 pfu, and changes of cell morphology were observed after 48 hours. Note that the cells expressing truncated myocilin became large and round.

Figure 7.

 

Diminished proliferation of TM cells expressing truncated myocilin. TM cells in 96-well plates were infected for 2 hours with each viral vector carrying truncated myocilin-GFP (mt MYOC), wild-type myocilin-GFP (wt MYOC), stromelysin-GFP (stromelysin), and GFP at the MOI indicated and cultured further for 72 hours. Plotted data are the mean ± SD (n = 3).