Mediation of Laser Trabeculoplasty–Induced Matrix Metalloproteinase Expression by IL-1β and TNFα

Abstract

purpose. Laser trabeculoplasty of the anterior uveal region of the trabecular meshwork induces sustained matrix metalloproteinase expression within the juxtacanalicular region of the meshwork. Studies were conducted to test the hypothesis that a factor mediates this response and to identify the factor.

methods. Human anterior segment organ cultures were subjected to laser treatment using standard clinical parameters and were returned to culture for 8 hours. The resultant 8-hour–conditioned culture medium was then tested for factor activity by evaluating its ability to produce two typical trabecular responses to laser treatment, that is, to induce stromelysin expression or to trigger cell division, when applied to fresh organ cultures or to cell cultures. Confocal immunohistochemistry of the laser-treated organ cultures and western immunoblot analysis of the conditioned medium were used to evaluate changes in potential candidates for the factor activity. The ability of the interleukin (IL)-1 receptor antagonist (IL-1ra)– and of tumor necrosis factor alpha (TNFα)–blocking antibodies to eliminate the stromelysin induction was evaluated.

results. Medium conditioned for 8 hours induced typical trabecular cell division in anterior segment organ cultures. Medium conditioned for 8 hours, but not for 30 minutes, induced typical increases in stromelysin expression in these organ cultures and in cell cultures. After 8 hours, both trabecular cells in laser-treated organ cultures and in the conditioned medium contained elevated levels of IL-1β and TNFα. The laser-treated organ cultures contained elevated levels of IL-1α, but it was not secreted into the medium. The ability of conditioned media to induce stromelysin expression was partially blocked by either the IL-1ra– or the TNFα–blocking antibody.

conclusions. Laser trabeculoplasty induces the expression and secretion of both IL-1β and TNFα within the first 8 hours after treatment. These cytokines then mediate increased trabecular stromelysin expression. Putatively, this initiates remodeling of the juxtacanalicular extracellular matrix, a likely site for the aqueous outflow resistance, and thus restores normal outflow facility.

Glaucoma is a major blinding disease. Between 1.5 and 2.47 million Americans and as many as 66.8 million persons worldwide are estimated to have glaucoma. ^{1 2 3 4} Open-angle glaucoma appears to be a family of disease entities, which produce a typical pattern of visual field defects. ¹ A major risk factor for this glaucomatous optic neuropathy is elevated intraocular pressure (IOP), caused by increased resistance to aqueous humor outflow localized within the trabecular meshwork (TM). ^{1 2 5 6 7}  

Laser trabeculoplasty (LTP), focal laser photocoagulation of the TM, remains a common tool for managing glaucoma. ^{1 7 8} A typical treatment protocol is to apply about 50 uniformly spaced 50-μm diameter burns to 180° of the anterior-central TM with an argon dye laser set on blue-green mode (predominantly 488–514.5 nm) with 0.1-second duration at a power setting of 0.7 to 1.0 W. ⁸ A number of variations of this treatment are used, and a wide variety of new laser or related methodologies are available or are in developmental stages. The parameters listed above were derived empirically, primarily from clinical observation. ^{8 9 10 11 12} The success rate of LTP is moderately high (40%–85%), depending on the patient sample and criteria used. ^{1 8 13 14 15} The usual response to LTP is a transient elevation of IOP, during which maximum tolerated medication is given, followed by an IOP reduction within 2 to 4 weeks. ^{8 9 12 16 17} The initial elevation of IOP after LTP may be due to trabecular obstruction by cellular and ECM debris, pigment granules, and perhaps some inflammatory effects. ^{8 9 12} The size of the eventual IOP reduction is roughly proportional to the extent of the pretreatment IOP elevation. ^{8 9 16 18} LTP treatment is not usually permanent, ^{9 19} which suggests that it is correcting a consequence of, but not curing, the disease. The efficacy of retreatment protocols is variable. ^{9 20 21 22}  

Within the first 24 to 48 hours, LTP triggers increased trabecular cell division ^{23 24 25 26} that is localized in humans predominantly to the trabecular insert, a nonfiltering region anterior to Schlemm’s canal and beneath Schwalbe’s line. ²⁷ Within 2 weeks after LTP, many of these newly divided cells appear to have migrated from the trabecular insert to repopulate and repair the LTP burn sites. ²⁷ Previous studies have shown that significant remodeling occurs in the trabecular juxtacanalicular ECM and possibly in the structural organization of the tissue. ^{17 28 29 30 31 32 33 34 35 36 37 38}  

Trabecular induction of several members of the matrix metalloproteinase family, particularly stromelysin and gelatinase B, in response to LTP ³⁹ may account for this remodeling. This induction is particularly prevalent, within the juxtacanalicular region of the meshwork. ⁴⁰ This family of secreted ECM-degrading enzymes, the matrix metalloproteinases, ^{41 42 43 44 45 46} are thought to be responsible for normal turnover and maintenance of the trabecular ECM. ^{47 48 49} Manipulation of these enzymes’ activity reversibly modulates aqueous outflow facility in perfused human anterior segment organ culture. ⁵⁰ The trabecular expression of these metalloproteinases and their tissue inhibitors, the TIMPs, is increased by a variety of growth factors and cytokines, particularly IL-1 and TNF. ^{49 51 52}  

The site of the LTP burns is physically remote from the site of stromelysin induction. The burns are in the anterior or central portion of the uveal meshwork and rarely penetrate more than 1/3 of its depth. The cell division is primarily in the trabecular insert, and the stromelysin induction occurs primarily in the juxtacanalicular region. Clinical treatment of only 180° of the meshwork is approximately as effective as treatment of the full 360°. We found that treatment of only 180° triggers both cell division and stromelysin induction throughout the full 360° of the meshwork. These observations lead us to hypothesize that these LTP responses might be mediated by a factor that is released in response to treatment. To test this hypothesis, we evaluated the ability of culture medium, conditioned for 8 hours by laser-treated organ cultures, to trigger these responses on trabecular cells and on untreated organ cultures. We then evaluated the possibility that the mediating factor was one of several growth factors or cytokines that we had shown to induce trabecular matrix metalloproteinase expression. ^{49 51 52} Two candidates, IL-1β and TNFα, were strongly induced by LTP treatment. The IL-1ra– and TNFα–blocking antibodies significantly reduced the ability of LTP-conditioned medium to induce stromelysin. 

Materials and Methods

Materials

Stromelysin antibodies were obtained from Triple-point Biologicals (Portland, OR); recombinant human IL-1α, IL-1β, IL-1ra, and TNFα and IL-1α and IL-1β antibodies were from R&D Systems (Minneapolis, MN); the TNFα antibody was from BioSource International (Camarillo, CA); horseradish peroxidase- or FITC-conjugated secondary antibodies were from Sigma (St. Louis, MO) or Molecular Probes (Eugene, OR); human lung fibroblasts were from ATCC (HFL1 CCL 153; Rockville, MD); Dulbecco’s modified Eagle’s Medium (DMEM) and antibiotic/antimycotic solution were from Gibco BRL (Gaithersburg, MD); fetal calf serum was from HyClone (Logan, UT); chemiluminescent detection kits were from Pierce (Rockford, IL); PicoGreen DNA assay reagent was from Molecular Probes; RNA extraction kits were from QIAGEN (Valencia, CA); chondroitinase A was from Seikagaku America (St. Petersburg, FL). 

Culture, Laser Treatments, and Conditioned Media

Human eye bank eyes were obtained from the Lion’s Eye Bank of Oregon (Portland, OR) and were used for stationary anterior segment organ culture ^{27 53} or to initiate trabecular meshwork cell cultures. ^{49 51} Porcine eyes, obtained from Carlton Packing Company (Carlton, OR), also were used to initiate trabecular meshwork cell cultures for some studies. ⁵¹ Human or porcine trabecular meshwork cells and human lung fibroblasts were cultured in DMEM with 100 U/ml penicillin G, 100 μg/ml streptomycin sulfate, 0.25μ g/ml Fungizone (Gibco BRL), and 10% fetal calf serum at 37°C, 100% humidity, and 5% CO₂. Trabecular cells were used by passage 5 and were made serum-free for 48 hours before initiating and during experiments. Human anterior segments explants, comprised of the intact cornea, the undisturbed trabecular meshwork, and a 2- to 5-mm rim of sclera with the ciliary body and iris gently removed, were stabilized in stationary organ culture for 7 days before starting experiments. ⁵³ These explants were cultured serum-free in 6-well dishes under conditions as detailed above. For laser treatments, stabilized, paired-eye anterior segment explants were subjected to standard, clinical-parameter LTP- or sham treatment ^{23 24 27 31 39 40} and returned to culture for 8 hours to condition the culture medium. The LTP treatment protocol was to apply 50 uniformly spaced 50-μm diameter burns to 180° of the anterior-central TM with an argon dye laser set on blue-green mode with 0.1-second duration at a power setting of 1.0 W. For the sham-treatment, the explant was exposed to the exact same process, but no LTP burns were applied. After the medium was conditioned for 8 hours, it was collected and either analyzed immediately or aliquoted and stored at −20°C. 

Trabecular [³H]Thymidine Incorporation

[³H]Thymidine uptake by trabecular cells in explants was evaluated as detailed earlier. ^{23 24 27} Briefly, medium that had been conditioned for 8 hours by either LTP- or sham-treated explants was diluted 1:1 with fresh medium.[ ³H]Thymidine (3 μCi/ml) was added and the medium was incubated with stabilized untreated paired-eye explants for 48 hours. The explants were then rinsed, bisected, fixed, and processed for autoradiography as previously detailed. ²³ Each microscope slide contained 16 sections (3 μm thick) taken from one quadrant of an explant, avoiding serial sections, which could contain portions of the same cell or nucleus. Eight pairs of eyes were exposed to conditioned medium and analyzed; from each eye, four slides (one from each quadrant) with 16 nonserial sections per slide were analyzed by two different individuals. Because each section contains about 100 to 125 trabecular cells, over 50,000 cells were counted for each of the LTP and the sham treatment groups. Using a masked protocol, nuclei with more than 20 autoradiographic grains were identified, and their position within the meshwork was determined. The four regions selected were the trabecular insert, the trabecular lining of Schlemm’s canal, the posterior meshwork, and the central meshwork. The total number of nuclei in each section also was determined. The data are presented both as the percentage of total cells in the respective section that exhibit more than 20 grains over their nuclei (total group) and as the percentage of the total cells in the section that were positive within each region. 

LTP Factor Assay

To assay for an LTP factor, medium that had been conditioned for 8 hours by either LTP- or sham-treated explants was diluted as indicated with fresh culture medium or was added without dilution to either untreated anterior segment organ cultures or to 12- or 24-well plates of densely confluent human or porcine trabecular meshwork cells or human lung fibroblast cells. After 24, 36, or 48 hours, the culture medium was removed, and stromelysin levels were determined by western immunoblot analysis or by zymography. Microscopic cell counts or PicoGreen DNA assays following the manufacturer’s instructions (Molecular Probes) was used in some cases to establish that the number of cells per well did not vary significantly with treatments or between wells because of plating differences. 

Confocal Immunohistochemistry

Eight hours after sham or LTP treatment, explants were rinsed in phosphate-buffered saline (PBS), and two wedges from opposite sides of the explants were removed and embedded in OCT containing 2.5% glycerol, quick-frozen, and stored at −80°C. Sections (6 μm) were cut with a digital microscope (1720 Cryostat; Leitz, Leica, Germany), using standard methods. Briefly, sections were thaw-mounted onto Superfrost Plus (Fischer Scientific, Pittsburgh, PA) slides, immersed in cold acetone for 2 seconds, and stored at −80°C. Slides were warmed to room temperature, fixed in 4% paraformaldehyde in PBS for 10 minutes, and then rinsed briefly in PBS. After preequilibration for 5 minutes at 37°C in buffer, 0.1 U proteinase-free chondroitinase A was added per milliliter and incubated for 8 minutes at 37°C. ^{53 54} Slides were washed in room temperature TBS (50 mM Tris, 150 mM NaCl, pH 7.4) and incubated with blocking buffer (TBS with 0.3% Triton X-100) for 30 minutes. Primary antibody (2–5 μg/ml) in blocking buffer was incubated with slides for 1 hour at 100% humidity, and slides were rinsed in TBS twice for 3 minutes each time and incubated with secondary antibody (0.8–2.5μ g/ml) in blocking buffer. Fluorescein-conjugated antibody incubations were for 45 minutes in the dark at 100% humidity. Sections were rinsed twice for 5 minutes each in TBS and equilibrated with antifade buffer (Molecular Probes) for 5 minutes Slides were coverslipped, sealed, and stored in the dark at 4°C until analysis. 

Confocal microscopy was conducted as detailed earlier. ^{50 55 56} Each antibody was evaluated in at least three separate experiments, and a total of 28 pairs of eyes were analyzed. Addition controls included sections from sham and laser treatment without primary antibody. Confocal instrument gain and zero settings were optimized below saturation on an intensely staining sample, and then no setting changes were made for the complete set; images were processed together to avoid introducing artificial differences. 

Western Immunoblot Analysis and Zymograms

Western immunoblot analysis and zymography were conducted as previously detailed, ^{49 51 57} except that horseradish peroxidase–conjugated secondary antibodies and chemiluminescent detection (Super Signal; Pierce) were used following the manufacturer’s instructions. For zymography, the medium was used without concentration; for western immunoblot analysis, the medium was concentrated 10 or 20 times using 10-kDa limit, Centricon spin-concentrators (Amicon, Beverly, MA). In either case, volumes were adjusted before applying to gels to normalize for possible processing differences between samples. Sample normalization was on a “per well of cells” basis, where cell numbers or cellular DNA content had been shown to be consistent between wells, as discussed earlier, or on a“ per explant” basis to avoid skewing due to changes in total media protein with treatments. 

RNA Isolations and Dot Blots

RNA was isolated as previously described ^{40 58} or using RNeasy kits as directed by the manufacturer (QIAGEN). Methods for dot blots were as described earlier. ⁴⁰ The total cellular RNA from four trabecular meshworks was pooled. Sample application for dot blots was normalized, based on total RNA extracted, as estimated from the absorbance at 260 nm; in most experiments, several dilutions of each sample were added to adjacent wells of the dot blot apparatus to established linearity ranges. Treatment of aliquots with RNase-free DNase before blotting did not affect results, whereas pretreatment with DNase-free RNase eliminated the dots. 

Results

LTP-Conditioned Medium Effects on Trabecular Cell Division

Because we had previously demonstrated that LTP treatment triggered a distinctive pattern of trabecular cell division, ^{23 27} the possible effect of an LTP-induced factor on cell division was evaluated. Media, conditioned for 8 hours by sham- or LTP-treated explant pairs from the same donor, were diluted 1:1 with fresh media and then applied to untreated explants pairs.[ ³H]Thymidine was added to the medium, and explants were cultured for 48 hours. The percentage of total trabecular nuclei exhibiting [³H]thymidine label in either the insert or in the total explant was significantly higher when treated with LTP- than with sham-conditioned medium (P < 0.0001 and P < 0.01, respectively; Fig. 1 ). The percentage of positive labeling nuclei in the region of Schlemm’s canal, the posterior meshwork, or the central meshwork was not significantly different between treatments. 

LTP- and Sham-Conditioned Media Effects on Trabecular Stromelysin mRNA and Protein Expression

Because trabecular stromelysin expression is the most well-defined molecular response to LTP treatment, ^{39 40} the possibility of an LTP-induced factor mediating this response was analyzed. Exposure of fresh explants to media that had been conditioned by LTP- or sham-treated explants for 8 hours resulted in increases in trabecular stromelysin mRNA levels in the fresh explants (Fig. 2) . At various times after addition of conditioned medium, explants were removed, and trabecular total RNA was extracted, applied to dot blots, and probed to determine stromelysin transcript levels. An autoradiograph from one typical experiment is shown in Figure 2A , and results of densitometric scans of another experiment are seen in Figure 2B . Four separate experiments with 3 to 12 pairs of explants in each were conducted. Maximum increases of two- to fourfold in stromelysin mRNA were seen between 8 and 24 hours in each of these experiments. 

Exposure to media that had been conditioned by LTP-treated explants increased stromelysin and gelatinase B but not gelatinase A protein levels compared to the effects of media conditioned by sham-treated explants (Fig. 3A ). Media that had been conditioned for 8 hours by LTP- or by sham-treated explants was diluted 1:1 with fresh media and incubated for 48 hours with human trabecular meshwork cells. In parallel as a positive control, explants that had been LTP-treated were left in their own media for 48 hours before analysis (LTP direct). The levels of these MMPs in the 8-hour–conditioned media were sufficiently low that additional media concentration steps were necessary to detect them (data not shown). Essentially no pro-gelatinase B (∼92 kDa) or activated gelatinase B (∼88 kDa) and only a small amount of activated and truncated gelatinase B (∼67 kDa) are produced by trabecular cells exposed to the sham-treatment media. LTP direct or LTP media both induce dramatic increases in pro-gelatinase B and strong increases in the other two forms. Pro-gelatinase A (∼72 kDa) is not appreciably effected by LTP treatment. Pro-stromelysin is dramatically increased in response to direct LTP- or to LTP treatment–conditioned media, relative to sham-treatment–conditioned media (Fig. 3A) . 

In separate but similar experiments (Fig. 3B 3C) , media was conditioned for 8 hours by LTP-treated explants; diluted 1:1, 1:5, or 1:10 with fresh media; and then incubated with human trabecular meshwork cell cultures for 48 hours. This was compared to the effect of media conditioned by sham-treated explants that were diluted 1:1 and by fresh media (Fig. 3C) and added to parallel cells for 48 hours. Stromelysin western immunoblots (Fig. 3B) and gelatinase B zymogram (Fig. 3C) band densities were determined by densitometric scanning. Stromelysin and gelatinase B indiction is dramatic in response to exposure to 1:1 diluted LTP- but not sham-treated explant-conditioned media. Greater dilutions, such as 1:5 or 1:10, were minimally effective. Gelatinase A levels remained unchanged by all these treatments (data not shown). 

We also determined whether the factor activity was heat-labile. Heating the LTP-conditioned medium for 1 hour at 80°C reduced its ability to stimulate stromelysin when added to fresh explants, by approximately 80%. Trabecular cells exposed to sham-conditioned medium or to sham-conditioned and then heated medium produced approximately the same low levels of stromelysin (not shown). The effects of these media on gelatinase B expression by human trabecular cells was similar to the effects on stromelysin, and gelatinase A levels were not affected either (data not shown). 

To determine whether the factor is released or produced and secreted in response to LTP treatment, we dissected untreated TMs from explants and subjected them to hypotonic shock, sonication, or Triton X-100 extraction. Equivalent amounts of these extracts were unable to induce stromelysin expression by trabecular cells (not shown). We also subjected paired explants to LTP and sham treatment and allowed them to condition medium for only 30 minutes. This conditioned medium also was unable to induce stromelysin expression, when applied to untreated explants or trabecular cells (not shown). 

LTP-Treatment Effects on Cellular IL-1α, IL-1β, and TNFα Levels and Media IL-1β and TNFα Levels

Confocal immunohistochemistry of anterior segment explants 8 hours after LTP treatment shows strong LTP-induced increases in the levels of IL-1α, IL-1β, and TNFα but no appreciable changes in platelet-derived growth factor (PDGF) bb levels (Fig. 4A ). Some basal immunostaining is apparent in sham-treated controls. Controls in which the first antibody was omitted showed lower levels of nonspecific staining (not shown). At higher magnification (Fig. 4B) , the LTP-induced increases in IL-1α and -1β and TNFα appear to be predominantly punctate and cytoplasmic. Some extracellular matrix immunostaining cannot be ruled out by these methods, although it is clearly not the dominant site of the immunostaining. 

Comparison of western immunoblot analysis of the 8-hour sham- and LTP-treated explant-conditioned media (Fig. 5A ) shows no detectable IL-1α or PDBFbb with or without treatment. However, large increases are apparent in IL-1β and TNFα levels in response to LTP treatment. Bands are at predicted sizes for the mature peptides, based on the literature and recombinant protein standards were included for the two that were not detectable, that is, IL-1α and PDGFbb. Because the immunostaining for IL-1α was very definite, without appreciable IL-1α in the medium, trabecular cells from LTP- and sham-treated explants were extracted with Triton X-100, followed by extraction with sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) sample buffer. Analysis of these extracts by western immunoblot analysis (Fig. 4C) shows that cellular IL-1α is elevated by LTP treatment. Triton did not release the IL-1α, whereas the sample buffer, containing SDS, did. Detectable basal levels in the sham-treated explants also are present. In both the sham- and LTP-treated trabecular cells, IL-1α is observed as a doublet at approximately 31 kDa, which is the size of the pro-form of this cytokine and not at the size of the mature peptide form. Thus, unprocessed IL-1α is present in trabecular cells and is increased by LTP treatment but is not secreted nor loosely associated with the plasma membrane. 

IL-1ra– or TNFα-Blocking Antibody Effects on Stromelysin Induction by LTP Medium

When IL-1ra is added to medium that had been conditioned for 8 hours by LTP-treated explants followed by assay of the medium for stromelysin induction by cell cultures (Fig. 6A 6B ), approximately 50% of the LTP-factor activity was blocked. This was a statistically significant difference, when compared to the same medium without the receptor antagonist added (P < 0.008, n = 3 pairs of explants). Western immunoblots (Fig. 6A) and the results of densitometric scans (Fig. 6B) are shown for three different pairs of explants. The medium was concentrated 10-fold before application to gels for stromelysin western immunoblot analysis. The 8-hour–conditioned medium from sham and LTP treatments did not contain appreciable levels of stromelysin (not shown). The cells exposed to sham-conditioned medium have lower but detectable levels of stromelysin, and this is slightly but not significantly reduced by the receptor antagonist as well. Human fibroblasts were used as an assay here because we had previously shown that porcine trabecular meshwork cells exhibit a significantly lower sensitivity to human IL-1, whereas human fibroblasts and trabecular cells exhibit essentially identical behavior in terms of stromelysin induction ^{51 52} (unpublished observations). 

In several similar experiments, a polyclonal (PAB) and a monoclonal (MAB) blocking antibody to TNFα also were able to reduce factor activity, measured similarly as stromelysin levels produced by cell cultures in response to medium exposure with or without the addition of blocking antibodies as indicated (Fig. 6C 6D) . Parallel treatment with a PDGFbb-blocking antibody reduces stromelysin induction only slightly. The final culture medium was concentrated 20 times before addition to the gels, and the exposure time was 36 hours. 

Discussion

The working hypothesis that laser trabeculoplasty triggers the production and secretion of a factor, which then mediates the stromelysin and gelatinase B induction, is supported by these studies. IL-1β and TNFα both appear to contribute significantly to the LTP-factor activity, as evaluated by stromelysin induction. IL-1α increases also are triggered by LTP treatment, but IL-1α is not secreted and remains primarily in the pro-form within the trabecular cells. Thus, IL-1α is not itself the LTP factor. Using a different system and conditions to evaluate laser induction of growth factors or cytokines, a previous study ⁵⁹ also reported an increase in cellular but not secreted IL-1α. Of the array of growth factors and cytokines that we screened for their ability to stimulated trabecular matrix metalloproteinase expression, IL-1 and TNF were the most effective. ^{51 52} Although PDGFbb also stimulated trabecular metalloproteinase expression, ⁵¹ it is not induced by LTP treatment and the PDGFbb-blocking antibody has minimal effects on LTP-factor activity. We have not ruled out additional contributions to the factor activity by other cytokines or growth factors. 

The effects of LTP- but not sham-treated explant-conditioned medium on trabecular cell division (Fig. 1) mimics the effects previously reported for direct LTP. ^{23 27} This provides evidence for the release of a factor in response to LTP. The factor initiating this cell division could be the same as or different from the factors that trigger the trabecular MMP response. This response is localized primarily to the trabecular insert, as previously reported, ²⁷ and is of a similar magnitude, with 0.2% to 0.3% of total trabecular cells dividing within a 48-hour period and LTP inducing somewhat greater than a doubling in the basal rate. ^{23 27} On the basis our earlier study, ²⁷ we believe that in humans only the “stemlike” trabecular insert cells are capable of significant division rates in vivo. That study provided evidence that these freshly divided insert cells then migrate out into the meshwork and repopulate areas where TM cells have been lost. However, the role of this phenomenon in the molecular mechanism of LTP as a treatment for glaucoma is not clear. ^{39 40} Although this is an intriguing question, it was not pursued in more detail herein. 

Because IL-1α and -1β and TNFα can each stimulate their own and each others’ expression, an amplification cascade seems feasible. ^{60 61 62 63 64} Several types of trabecular cell disruption or extraction do not release the factor activity; thus, the factor is most likely induced rather than released by LTP treatment. Although IL-1α is not secreted, it could play an initiating role in an LTP-factor amplification cascade, for example, by acting within the cells to stimulate IL-1β and/or TNFα expression and secretion. One could also evoke that cellular IL-1α is released from a zone of damaged trabecular cells near the burns. Because this would involve only a relatively small number of cells, the amount of IL-1α released may not be sufficient to be detected by western immunoblot analysis. However, this could be enough to initiate an amplification cascade of IL-1β and/or TNFα expression and secretion spreading out from the burns and eventually reaching a level to induce metalloproteinase expression in the juxtacanalicular region of the meshwork. ⁴⁰ Other molecular mechanisms could be responsible for initiating the cascade, such as heat shock of the cells near to but not directly damaged by the LTP burns. Thus, the importance of the cellular IL-1α increase is not clear. The molecular mechanism for the IL-1α, IL-1β, and TNFα induction by LTP also remains to be determined. 

We see very intense IL-1 and TNF immunostaining immediately surrounding individual LTP burns and an absence of similar immunostaining for PDGFbb (data not shown). This may suggest that the LTP effect is initially on a zone of trabecular cells near the burns. However, many cytokines and most antibodies bind nonspecifically to areas of damaged or disrupted tissue or extracellular matrix; thus, this increased localization at the site surrounding the LTP burns could easily be an artifact. 

We have previously shown that LTP treatment induces sustained stromelysin and gelatinase B expression, ³⁹ particularly within the juxtacanalicular region of the meshwork. ⁴⁰ We also have recently shown that increasing the trabecular matrix metalloproteinase activity reversibly increases outflow facility. ⁵⁰ Thus, the LTP-induced metalloproteinase increases and the juxtacanalicular extracellular matrix remodeling that this would putatively cause provide a plausible explanation for the efficacy of LTP treatment for glaucoma. Understanding at a molecular level how LTP ameliorates glaucomatous IOP elevations may allow the development of improved LTP treatment parameters. Interestingly, the Nd:YAG laser in either thermal or cutting mode also initiates these trabecular responses. ⁶⁵  

Although LTP is effective in many cases of glaucoma, it does not provide a permanent resolution and is not effective in all cases. Our observation, that the effects of LTP are mediated by a factor and the identification of this factor as IL-1β and TNFα may be useful in designing alternative therapies for glaucoma, would likely be facilitated by a detailed understanding of the signal transduction pathways involved in the induction of the factor(s) by LTP and in the induction of stromelysin and gelatinase B by these LTP factors. 

 Supported by the National Institutes of Health Grants EY08247, EY10572, EY03279, and EY07123, and by grants from the Glaucoma Research Foundation, American Health Assistance Foundation, Research to Prevent Blindness, and Alcon Laboratories.

 Submitted for publication March 11, 1999; revised August 3, 1999; accepted August 27, 1999.

 Commercial relationships policy: N.

 

Figure 1.

 

Effects of conditioned medium on trabecular cell division. Autoradiographic analysis of the percentage of trabecular cell nuclei exhibiting [³H]thymidine uptake after exposure to medium conditioned for 8 hours by sham- or LTP-treated explants is shown. Nuclei with more than 20 autoradiographic grains were counted as positive in the trabecular insert (Insert), Schlemm’s canal lining (SC), posterior portion of the meshwork (Post), central portion of the meshwork including uveal, corneoscleral, and juxtacanalicular regions (Central), and total trabecular meshwork (Total). Data are presented as the percentage of the total trabecular cell nuclei in a microscope section that had more than 20 grains over their nuclei within each of the four regions and as the total of all four regions. Data are the means ± SD for counts of total nuclei in more than 1000 sections from eight pairs of eyes. P values were determined by Student’s t-test.

Figure 2.

 

Time course of effects of LTP-conditioned media on trabecular stromelysin mRNA levels. Media, conditioned for 8 hours by sham- and LTP-treated explants, were diluted 1:1 with fresh media and applied to untreated explants for various times as indicated. Trabecular total RNA was then extracted and applied to dot blots (1 μg of total RNA per dot) and probed to determine trabecular stromelysin mRNA levels. (A) One typical autoradiograph of samples after treatment for 6, 8, 12, and 24 hours shows the effects of incubation with sham- or LTP-treated media. (B) Results are shown of densitometric analysis of autoradiographs from separate experiments similar to those shown in (A). A total of four similar experiments were conducted (see the Materials and Methods section).

Figure 3.

 

Effects of LTP-conditioned media on trabecular stromelysin and gelatinase protein levels from western immunoblot analysis and zymograms. (A) Media that was conditioned for 8 hours by sham- and LTP-treated explants were diluted 1:1 with fresh culture media and applied to trabecular cell cultures. After 48 hours, media gelatinase A and B levels were evaluated by gelatin zymogram or stromelysin protein levels by western immunoblot. The “LTP Direct” lane shows levels of stromelysin and the gelatinases that were in media that was left on the LTP-treated explant for 48 hours. (B, C) Results of densitometric scans of stromelysin western immunoblots or gelatin zymograms, respectively, from separate experiments. Explants were LTP- or sham-treated and allowed to condition media for 8 hours. The media were then removed and diluted as indicated with fresh media. The “C” lane shows the effect of fresh media that had not been conditioned. These diluted media were added to porcine trabecular meshwork cells for 48 hours and then analyzed for stromelysin (B) or gelatinase B (C) by western immunoblot analysis or gelatin zymography, respectively. The actual conditioned medium did not contain appreciable levels of stromelysin or gelatinase B (not shown). The latent pro-gelatinase B (92 kDa), activated gelatinase B (88 kDa), and the activated and truncated gelatinase B (68 kDa) bands are plotted together. The gelatinase A levels did not change significantly and thus are not shown.

Figure 4.

 

Confocal immunohistochemical analysis of select cytokine or growth factor expression induced by LTP. Paired explants were subjected to sham or LTP treatment and returned to culture for 8 hours. Explants pairs were then subjected to confocal immunohistochemical analysis using the antibodies as indicated. Arrows mark Schlemm’s canal and the trabecular meshwork is below it in all panels. (A) The full trabecular meshwork is shown at lower magnification; (B) a higher magnification view of the central deep corneoscleral and juxtacanalicular region of the meshwork.

Figure 5.

 

Western immunoblot analysis of cytokine or growth factor expression induced by LTP. Paired explants were subjected to sham or LTP treatment and returned to culture for 8 hours. (A) The conditioned media were then collected and analyzed for cytokine or growth factor levels by western immunoblot analysis as indicated. Positive standards (+Std) using the respective human recombinant protein were added for verification in cases where no immunostaining was detected in the conditioned media (IL-1α and PDGFbb); the sham and LTP (Laser) treatment lanes are as indicated above the blots. The arrowheads point to the bands, all of which are observed to migrate at the predicted size for the mature peptide. Media from two different pairs of explants probed for IL-1β are shown one over the other. The mature forms of IL-1, TNFα, and PDGFbb migrate at approximately 17, 19, and 25 kDa, respectively (arrows); the unprocessed form of TNFα runs at approximately 36 kDa (upper arrow). (B) Western immunoblot analysis of cellular extracts, collected at 8 hours after similar treatment, were probed for IL-1α. “Triton” lanes represents Triton X-100 extract of dissected trabecular meshwork and “SDS” lanes represent SDS-PAGE sample buffer extract after the Triton extraction. The pair of IL-1α bands are at ∼ 31 kDa, the size of the pro-peptide form.

Figure 6.

 

Effects of IL-1ra and of TNFα and PDBFbb blocking antibodies on stromelysin induction by conditioned medium. Medium conditioned for 8 hours by sham- and LTP-treated explants was assayed for stromelysin inductive capability with or without the addition of IL-1ra (5 μg/ml) (A, B); polyclonal blocking antibodies (PAB) to TNFα (0.1 μg/ml) or PDGFbb (3 μg/ml) (C); or to a different batch of polyclonal antibody to TNFα (0.75 μg/ml) (D), as indicated above the lanes. Conditioned medium was incubated for 36 hours with human lung fibroblasts (A) through (C) or porcine trabecular meshwork cells (D); the cell culture medium was then assayed for stromelysin levels by western immunoblot analysis. Three pairs of explants were evaluated (A) and densitometric scans of these immunoblots (B) are shown. P < 0.008 comparing LTP medium with and without IL-1ra added. (C) Similar to (A) except that either a polyclonal blocking PDGFbb or TNFα antibody was added to the conditioned medium as indicated, before incubation. Medium was tested for stromelysin levels by western immunoblot analysis after exposure to cells for 36 hours. (D) LTP, sham, or new (unconditioned) media with or without a TNFα-blocking antibody was added to cells for 36 hours, and then the cell culture medium was assayed for stromelysin protein levels.

The authors thank the Lion’s Eye Bank of Oregon (Portland, OR) for providing donor eyes; and the Microbiology and Molecular Immunology Core Facility (Oregon Health Sciences University, Portland, OR) for confocal microscopy.