Ultrastructural and Biochemical Evaluation of the Porcine Anterior Chamber Perfusion Model

Abstract

purpose. To evaluate a porcine anterior chamber perfusion model and to test the transferability of data obtained with this model to the human system.

methods. Porcine eyes were obtained from a local abattoir and processed within 2 hours after death. Anterior chambers of 42 pairs of eyes were dissected with removal of lens, vitreous, iris, and ciliary processes and perfused for 72 (40 pairs) or 140 (2 pairs) hours with medium or medium supplemented with 10 ng/mL transforming growth factor (TGF)-β₂. Facility was continuously measured. Afterward, trabecular meshwork (TM) specimens from all quadrants were prepared, and sections were analyzed morphologically and with immunohistochemical methods. TM sections of 10 nonperfused pairs of eyes were used as the control. RNA and protein was extracted from the TM specimens. Expression of αB-crystallin, fibronectin (FN), plasminogen activator inhibitor (PAI)-1, thrombospondin (TSP)-1, and connective tissue growth factor (CTGF) mRNA and protein in medium-perfused and TGF-β₂-perfused anterior segments was examined by Northern and Western blot analyses.

results. The nonperfused TM showed prominent differences between the temporal and nasal quadrants. Temporally, the ciliary muscle (CM) was pronounced, the scleral sulcus was long and flat, and the scleral spur extended toward the iris root. Nasally, the CM was thin, the sulcus deep, and the spur compact. The outer TM was expanded between the scleral spur and cornea throughout the entire circumference. On the ultrastructural level, the elastic network was connected to the cribriform TM cells and the aqueous plexus endothelium. Perfusion itself had only small effects on the morphology of the outer TM. Aqueous plexus loops remained open, and TM cells showed no signs of necrosis or pyknosis. αB-crystallin expression was significantly increased in perfused eyes. Perfusion with TGF-β₂ for 72 hours reduced outflow facility to approximately 60% of that of the medium-perfused control. TM cells adjacent to putative drainage pathways showed enlarged cisterns of rough endoplasmic reticulum (rER), a sign of active protein synthesis. Expression of αB-crystallin and FN mRNA were elevated by factors of 5 and 3, respectively. The proteins were upregulated by a factor of 2.5. In addition, TGF-β₂ upregulated PAI-1 (1.7-fold) and TSP-1 (1.6-fold) proteins, two factors shown to be TGF-β₂ responsive in human TM cell culture experiments. CTGF expression was not altered.

conclusions. These new ultrastructural investigations indicate that the cribriform and subendothelial regions of the porcine TM have an architecture similar to that of the primate TM. The biochemical and physiological response to TGF-β₂ was identical with that described in human TM cell culture and anterior chamber perfusion. The porcine anterior chamber perfusion model is valid for the human system.

It is well established that in most glaucomatous eyes, intraocular pressure (IOP) is elevated as a result of an increase in the aqueous humor (AH) outflow resistance. Comparative morphologic and physiological studies, ¹ as well as direct measurements, ^{2 3} have indicated that approximately 90% of the resistance is located within the cribriform or juxtacanalicular region of Schlemm’s canal (SC). In eyes with primary open-angle glaucoma (POAG) there is an increase in outflow resistance and a significant increase in the amount of the so-called sheath-derived (SD) plaque material. ^{4 5} In contrast, measurements of putative draining pathways from morphologic sections suggest that the remaining optically clear spaces within the cribriform region are still large enough for a normal outflow facility, ⁶ from which it was concluded that there is additional material besides the SD plaques contributing to outflow resistance. This material is presumably lost during the fixation and embedding process for electron microscopy and therefore is not educible when using this technique. The composition of the material has not yet been completely characterized because of this loss and also because the factors responsible for an increased synthesis of that material by trabecular meshwork (TM) cells have not been completely identified to date. 

One of the factors that presumably plays a role in glaucoma manifestation is TGF-β₂. Studies have demonstrated that approximately 50% of patients with POAG have significantly increased levels of this cytokine in the AH. ^{7 8 9 10} Based on this, numerous studies have been undertaken to elucidate the effects of TGF-β₂ with respect to extracellular matrix (ECM) modifications in the TM, using human TM cell cultures. ^{11 12} Physiological studies, using the human anterior chamber (AC) perfusion system, have shown that perfusion with TGF-β₂ significantly reduces outflow facility, ¹³ and ultrastructural evaluation of the perfused material has revealed an increase of ECM underneath the inner wall of SC. Meanwhile, biochemical and molecular in vitro investigations have shown that TGF-β₂ is a potent activator of ECM component synthesis as well as of the expression of ECM degradation regulators. Among the proteins shown to be TGF-β₂ responsive were fibronectin (FN), an integral constituent of basement membranes (BM), tissue-type transglutaminase (tTgase), an enzyme cross-linking ECM to undegradable complexes ¹¹ and plasminogen activator inhibitor (PAI)-1, a factor inhibiting matrix metalloproteinase (MMP) activation. ^{11 12} These data indicate that TGF-β₂ has the ability to activate the synthesis of ECM material and inhibit ECM degradation via PAI-1, resulting in an accumulation of ECM. However, to get insights into the composition of this material and the regulatory mechanisms and pathways involved, as well as to test other possible glaucoma-promoting factors, further molecular and morphologic investigations are essential. 

Because human material is available in limited quantities only, researchers have used animal models. One frequently used model is the porcine eye, because it provides almost unlimited material and the variations in age compared with human donor eyes are relatively small. However, morphologic study of the porcine eye identified significant differences with respect to the overall anatomy and the architecture of the TM compared with human eyes. It was thus not clear to what extent data obtained using the porcine perfusion model can be applied to the human situation. The purpose of the present study was therefore to evaluate the applicability of the porcine perfusion model. For that, detailed morphologic analysis of the porcine outflow system and new ultrastructural investigations of the juxtacanalicular region were conducted. Moreover, perfusion studies were performed to test first the influence of the preparation technique on the outflow facility and second the effect of the perfusion itself on TM morphology. Finally, the response of the porcine TM to perfusion with TGF-β₂ was analyzed physiologically and biochemically. In Northern and Western blot studies we focused on factors known to be upregulated by TGF-β₂ in perfused human eyes and human TM cell cultures. 

Materials and Methods

Preparation of Porcine Eyes

Porcine eyes were obtained from a local abattoir and processed within 2 hours after death. Eyes not designated for perfusion studies (10 pairs) were equatorially sectioned, and the lens and vitreous were removed. Posterior segments were discarded, and 1- to 2-mm wide specimens of all four quadrants of the anterior segments were excised and subsequently fixed in either 4% paraformaldehyde (PFA) for 4 hours for immunohistochemical investigations (four pairs) or in Ito’s fixative for at least 12 hours for light and electron microscopy (six pairs). 

The preparation technique for eyes used in perfusion experiments was a modification of that described for human eyes by Johnson et al. ^{14 15} Eyes were equatorially sectioned and vitreous, retina, and choroid were removed (Fig. 1A) . The posterior lens capsule was incised crosswise, and the lens proper was removed, leaving the lens capsule in place. In initial studies, we found that the success rate was absolutely dependent on the amount of pigment set free during the preparation. Excessive pigment dispersion into the TM caused TM cell degeneration and rendered the anterior chamber preparations unusable for perfusion studies. To minimize pigment dispersion, it was necessary to keep the lens capsule in place, as it prevented pigment infiltration into the TM during the further preparation. To remove iris and ciliary body (CB), a sagittal incision was made from the iris toward the CB at one place and the released pigment was immediately rinsed off with medium (Fig. 1B) . From that incision, the iris and CB were lifted with forceps and gently pulled outward, to apply a weak tension while the disconnection from the subjacent scleral spur (SP) and TM was performed with fine scissors (Fig. 1B) . With this method, the longitudinal portion of the CM and SP remained in place, visible as a ring of pigmented tissue (Figs. 1B 1C) . After extensive washes with Dulbecco’s modified Eagle’s medium (DMEM) to remove pigment and cell remnants, the anterior segments were placed in a perfusion chamber, perfused with DMEM, supplemented with 1.5 mg/mL glucose, 1% (vol/vol) fetal calf serum (FCS) and antibiotics (100 U/mL penicillin, 0.1 mg/mL streptomycin, and 17 μg/mL gentamicin; all from Invitrogen, Karlsruhe, Germany) at a constant flow rate of 4.5 μL/min. The anterior eye segments were maintained in an incubator (HeraCell; Kendro, Langenselbold, Germany) at 37°C in 7% CO₂, while the IOP was continuously monitored and recorded with a computer system. 

Perfusion Studies

After an adaptation period of at least 40 hours, medium perfusing the left anterior segment was changed to DMEM containing 10 ng/mL recombinant TGF-β₂ (Roche, Basel, Switzerland). The contralateral control received a similar medium exchange using DMEM without TGF-β₂. Forty pairs were perfused for a further 72 hours under these conditions. In two pairs, perfusion time was extended to 140 hours after solution exchange. At the end of the perfusion, 1- to 2-mm wide specimens of the four eye quadrants were excised. Specimens of 10 eye pairs were fixed in 4% PFA for immunohistochemical investigations, and the specimens of the other 30 eye pairs were fixed in Ito’s fixative for light and electron microscopy. 

From the remaining anterior eye segments, TM specimens were excised similarly as described elsewhere for human eye preparations. ¹⁶ In brief, the posterior border of the TM, adjacent to the pigmented borderline of the SP and the remnants of the longitudinal CM, was perforated with a scalpel. In a distance of ∼2 mm toward the cornea, a second perforation was set defining the anterior border of the TM (Fig. 1C) . This width of 2 mm guaranteed that the complete TM was demarcated throughout the entire circumference, although there were regional differences in TM width between quadrants. The TM was then removed as a complete band from the neighboring tissues with forceps (Fig. 1D) . Specimens were cryoconserved in liquid nitrogen for subsequent biochemical investigations. 

Only eyes with an intact morphology of the outer TM (i.e., open vessel loops and viable remaining outer TM cells) were used for statistical evaluation of perfusion data. In the present study, 32 of the 42 pairs (30/40 perfused for 72 hours, 2/2 perfused for 140 hours) had acceptable morphology and were considered for further analysis. The facility data are presented in two ways. One is by averaging the raw facility data in all eyes of one perfusion group, whereas the second is by normalizing the facility rates of each eye and then averaging the normalized data of all eyes in one perfusion group. Normalization was achieved by dividing the measured facility for a given eye at each moment by the average facility reading of that eye in a 10-hour period before medium exchange. The statistical significance of facility changes was computed from a paired two-tailed Student’s t-test, by using the raw or normalized facilities for the 10-hour period beginning 50 hours after medium exchange. 

Light and Electron Microscopy

The Ito-fixed sections were rinsed in cacodylate buffer for 12 hours. Thereafter, the specimens were postfixed in 1% OsO₄, dehydrated in an ascending series of alcohols, and embedded in Epon (Roth, Karlsruhe, Germany) according to standard protocols. Semithin sections were cut with a microtome (Ultracut OmU3; Reichert, Vienna, Austria) and stained with toluidine blue. Ultrathin sections were stained with uranylacetate and lead citrate and viewed with an electron microscope (EM 902; Carl Zeiss, Oberkochen, Germany). From each quadrant of the eye 20 semithin sections at least one ultrathin section was analyzed. 

Immunohistochemistry

For immunohistochemical studies, the PFA-fixed specimens of 10 perfused and 4 nonperfused pairs of anterior eye segments were washed thoroughly with Tris-buffered saline (TBS, pH 7.4) and placed in embedding medium (Tissue-Tek-OCT; Jung, Nussloch, Germany), frozen at −20°C and cut into 12-μm thin slices with a cryotome (Kryostat 1720; Leica, Bensheim, Germany). The sections were blocked for 1 hour at room temperature in blocking reagent (5% [wt/vol] nonfat dry milk powder in TBS; Blotto, Santa Cruz Biotechnology, Santa Cruz, CA), rinsed, and washed once for 10 minutes in TBS/0.3% (vol/vol) Tween-20 (TBST), and incubated with primary antibody diluted in 1.5% blocking agent overnight (Table 1) . The slices were then washed in TBS and incubated with fluorescein-conjugated secondary antibody diluted in TBST (Table 1)for 75 minutes at 4°C. 

Biochemical Analysis

Only specimens of eyes with intact morphology were used for subsequent biochemical experiments. Specimens of 16 of the 30 perfused eye pairs that fulfilled the morphologic criteria were used for RNA isolation and subsequent Northern blot analysis, the remaining specimens of 14 eye pairs were used for protein extraction and subsequent Western blot analysis. To obtain sufficient amounts of RNA and protein specimens of four eyes of each treatment group (TGF-β₂- and medium-perfused) were pooled. 

RNA Isolation and Northern Blot Analysis

Total RNA was phenol-chloroform extracted from pooled frozen TM specimens (TRIzol; Invitrogen). Structural integrity of the RNA samples was confirmed by electrophoresis in 1% Tris-acetate-EDTA (TAE)-agarose gels. Yield and purity were determined photometrically. For Northern blot analysis, 10 μg of total RNA was size fractionated by gel electrophoresis in 1% agarose gels containing 2.2 M formaldehyde, subsequently transferred onto a nylon membrane (Roche) by vacuum blotting and cross-linked at 1600 μJ (Stratalinker; Stratagene, La Jolla, CA). Membranes were prehybridized for 1 hour (Dig Easy Hyb solution; Roche) at 68°C. Hybridization followed overnight at 68°C in prehybridization solution containing 50 ng/mL of a specific antisense probe. To obtain gene-specific cDNA fragments, primer pairs for FN (forward 5′-gaagctctctctcagacaacca-3′, reverse 5′-aggtctgcggcagttgtcac-3′) ¹⁷ and αB-crystallin (forward 5′-caccacccctggatccg-3′, reverse 5′-cttctcttcacgggtgat-3′) ¹⁸ were used for PCR and products were subcloned into a commercial vector (pTOPO/TA; Invitrogen). Digoxigenin-uridine triphosphate (DIG-UTP)–labeled riboprobe synthesis was performed with a DIG-labeling kit (Roche) according to the manufacturer’s instructions. After hybridization, the membranes were washed twice in 2× SSC/0.1% SDS at room temperature, followed by two washes in 0.1% SDS for 15 minutes at room temperature and 68°C. Afterward, the membranes were rinsed, washed once for 5 minutes in washing buffer (100 mM maleic acid, 150 mM NaCl [pH 7.5], 0.3% Tween-20) and incubated for 1 hour in blocking solution (100 mM maleic acid, 150 mM NaCl [pH 7.5], and 1% blocking reagent; Roche). Alkaline-phosphatase conjugated anti-digoxigenin-antibody (α-DIG-AP; Roche) was diluted 1:10,000 in blocking solution and added to the membranes for 30 minutes. After incubation, the membranes were washed four times for 15 minutes each in washing buffer and equilibrated in chemiluminescence detection buffer (100 mM Tris-HCl and 100 mM NaCl [pH 9.5]) for 10 minutes. For chemiluminescence detection, the alkaline phosphatase substrate (CDP-Star; Roche) was diluted 1:100 in detection buffer and the membranes were incubated for 5 minutes at room temperature. After air-drying, the semidry membranes were sealed in plastic bags and chemiluminescence was detected (Lumi-Imager workstation; Roche, Mannheim, Germany). Exposure times ranged between 2 minutes and 1 hour. The quantification was performed on computer (Lumi-Analyst software; Roche). To assess the amount and quality of the RNA, 28S and 18S rRNA bands were visualized by methylene blue staining and images were taken (Lumi-Imager; Roche). Signal intensities were considered for quantification. Experiments were performed twice on each group of pooled eyes (n = 4). 

Protein Extraction and Western Blot Analysis

Pooled frozen TM probes were thawed and directly lyzed in RIPA lysis buffer (150 mM NaCl, 1% NP-40, 0.5% DOC, 0.1% SDS, and 50 mM Tris [pH 8.0]). Cell debris was removed by centrifugation for 5 minutes at maximum rpm at 4°C in a centrifuge (Eppendorf, Hamburg, Germany) and protein contents of the supernatants were photometrically determined by Bradford’s protein assay (Bio-Rad, Munich, Germany). All probes were supplemented with one third total volume 4× SDS-loading buffer, denatured by boiling for 5 minutes, and stored at −20°C until use. Samples of 5 μg protein were separated by SDS-polyacrylamide gel electrophoresis (PAGE) in separation gels from 6% to 15% at 25 mA constant, transferred onto nitrocellulose membranes (Protran BA 83, 0.2 μm; Schleicher & Schüll, Dassel, Germany) by tank blot at a constant 70 V for 45 to 60 minutes in 1× transfer buffer (10 mM CAPS [pH 11; 3-(cyclohexylamino)-1-propanesulfonic acid] 20% methanol, 0.1% SDS). Membranes were blocked in 5% blocking agent (Blotto; Santa Cruz Biotechnology) for 1 hour at room temperature under constant agitation. Primary antibodies were then added in the appropriate dilution in 1.5% blocking agent and allowed to react overnight under constant agitation at 4°C. After rinsing and washing once for 5 minutes at room temperature in TBST, the corresponding secondary antibodies were added in the appropriate dilution in 1.5% blocking agent, and blots were incubated for 30 minutes at room temperature. After three 10-minute washes in TBST and one wash for 10 minutes in detection buffer, alkaline phosphatase substrate (CDP-Star; Roche) diluted 1:100 in detection buffer was added, the membranes were incubated for 5 minutes at room temperature, and chemiluminescence signals were visualized (Lumi-Imager workstation; Roche). Exposure times ranged between 1 and 5 minutes. The quantification was performed on the workstation (Lumi-Analyst software; Roche). Experiments were performed twice on the tissue of each group of pooled probes (n = 3). 

Results

Histology and Electron Microscopy

Morphology of the Porcine TM.

The analysis of sagittal sections of normal, nondissected, nonperfused porcine eyes confirmed the descriptions published in previous studies. ^{19 20} There were pronounced regional differences in the circumference of the anterior segment, the most prominent observed between the temporal and nasal quadrant. In the temporal quadrant, the TM was expanded between the SP and cornea, but a significant portion toward the reticular TM was fibrously connected to the CM. The CM was pronounced and the scleral SP showed a prominent prolongation toward the iris root on the temporal side compared with the nasal side. The distance between SP and Descemet’s membrane was large, and the scleral sulcus appeared long and flat in the anterior–posterior dimension (Fig. 2A) . In contrast, in the nasal quadrant, the sulcus was deep, and the distance from SP to Descemet’s membrane was shorter than temporally. The SP had a compact appearance, whereas only a thin CM was visible. Nearly the entire TM was expanded between the SP and cornea, and the entire TM was thicker than in the temporal quadrant (Fig. 2B) . Despite these quadrant-specific differences of the architecture of the entire TM, the outer juxtacanalicular meshwork, into which the vessel loops of the outflow plexus extend, was expanded between SP and cornea and was morphologically similar within the entire circumference (Figs. 2A 2B) . 

Ultrathin sagittal and tangential sections through the juxtacanalicular, or cribriform, region revealed that in the porcine, like in primate eyes, the TM contained a three-dimensional elastic fiber system surrounded by collagen fibers and ground substance. The elastic fibers showed the same morphology as in the primate TM. They contained an electron-dense fiber core intermingled with electron-lucent material, and they were surrounded by a more or less homogeneous-appearing sheath material. From the sheath material, fine fibrils split and connected the elastic fibers to adjacent cribriform cells (Fig. 3A) . 

The TM cells in this region had a dendriform appearance, with long cytoplasmic processes through which they were in contact with each other. They showed patches of BM material and contained within the cytoplasm numerous ribosomes and rER (Figs. 3 4) . 

In the subendothelial region of the aqueous plexus, the elastic fibers were connected to the endothelial cells by sheath-derived fine fibrillar material (Fig. 4) . In addition to elastic fibers, the subendothelial region contained collagen fibers that showed a typical periodicity. This subendothelial extracellular material was interrupted by optically empty pathways facing BM-free areas of the endothelium of the vessel loops. The subendothelial cribriform cells often showed cisternae of rER filled with a homogeneous electron-lucent material (Fig. 4) . The endothelial cells had a discontinuous BM (Fig. 4)and in some of the BM-free areas, giant vacuoles within the endothelial cells were visible. 

Morphologic Changes after Perfusion.

Sagittal sections through perfused anterior segments, in which the iris and most of the CM had been removed, showed that the outer meshwork, into which the open vessel loops extended, was still expanded between the SP and cornea in all quadrants. In the temporal quadrant, parts of the inner meshwork were absent because of the preparation of the CB and also, in the other quadrants, the inner TM appeared slightly collapsed compared with the nonperfused eyes (Figs. 2C 2D) . Within the cribriform TM, however, there still were optically empty pathways directed toward the aqueous plexus. Numerous TM cells were present and showed no signs of pyknosis or necrosis (Figs. 2C 2D) . Open vessel loops were completely ensheathed by endothelial cells. At the ultrastructural level, there were no differences noted in the morphology of the juxtacanalicular region compared with that in the the nonperfused control eyes. This similarity was observed in all quadrants of the eyes. 

Immunohistochemistry of Normal and Perfused Porcine Anterior Chambers

Smooth-Muscle α-Actin.

Immunohistochemical staining for α-smA in normal, nonperfused control eyes and in medium-perfused anterior segments was identical. Single cells adjacent to the outflow loops and along the collecting vessels were labeled, whereas the TM cells proper remained unstained (data not shown). Similar labeling was observed in all quadrants. 

αB-Crystallin.

The staining pattern of αB-crystallin was significantly different in perfused anterior chambers compared with normal control eyes. In unperfused eyes, αB-crystallin labeling was restricted to TM cells of the inner, reticular TM and not seen in cells of the TM expanded between the SP and cornea (Fig. 5A) . In contrast, in perfused eyes, almost all TM cells showed an intense staining for αB-crystallin (Fig. 5B) . Quadrant-specific differences were not observed. 

Fibronectin.

Antibodies to FN stained extracellular material throughout the entire TM as well as adjacent to vessel loops and collector vessels. At the histologic level, slightly elevated levels of FN staining in media-perfused anterior chambers compared with unperfused controls were detected (Figs. 5D 5E) . The results were the same in all quadrants. 

Effect of TGF-β₂ Perfusion

Morphologic Changes.

The general morphology of the TM in TGF-β₂ perfused eyes was the same as described with the control perfusion. Despite this similarity in the overall morphology of the TM, there were distinct ultrastructural changes in the appearance of TM cells along the draining channels after TGF-β₂-treatment. The cells were enlarged and showed an increase in the size of the cisternae of the rER (Figs. 6A 6B) . This augmentation of cisternae of rER was already seen after 72 hours of perfusion. Prolonged perfusion for 140 hours did not result in a further increase of cell organelle profiles. 

Immunohistochemical Changes.

The overall staining patterns obtained by the use of α-smA antibodies on sections of TGF-β₂ perfused eyes was identical with those of media-perfused control anterior chambers (data not shown). The intensities of αB-crystallin and FN staining appeared pronounced in the sections of TGF-β₂-perfused eyes (Figs. 5C 5F) . However, the differences of staining intensities could not be adequately documented due to the already pronounced staining in the perfusion control eyes (Figs. 5B 5E) . Regional differences in the quadrants were not detectable. 

Biochemical Analysis

Northern and Western blot analyses of αB-crystallin and FN mRNA and protein, respectively, revealed that perfusion with TGF-β₂ for 72 hours significantly upregulated the expression of both (Fig. 7A) . By quantification of the Northern blot data, a 5.6 ± 0.7 (SD)-fold induction for the 0.8-kb αB-crystallin-specific mRNA and a 3.0 ± 0.3-fold induction for the 7.7-kb FN-specific mRNA was calculated compared with the levels in medium perfused anterior chambers (Fig. 7B) . Signal intensities of the 28S and 18S rRNA bands served as the control for equal loading and were considered for quantification. 

From the Western blot analysis (Fig. 8A) , inductions of 2.8 ± 0.6-fold for the 20-kDa αB-crystallin protein and of 2.6 ± 0.5-fold for the polymerized 240-kDa FN protein complex, respectively, were calculated (Fig. 8B) . Equal loading was assured by Bradford protein determination before SDS-PAGE, and equal transfer was tested by methylene blue staining of the membranes after chemiluminescence detection (not shown). Among the other proteins tested and known to be TGF-β₂ responsive in human anterior chamber perfusion or human TM cell culture, PAI-1 and TSP-1 showed a moderate but reproducible augmentation (Fig. 8A) . The 50-kDa PAI-1 protein was upregulated by 1.7 ± 0.1-fold and the 180-kDa TSP-1 protein by 1.6 ± 0.1-fold (Fig. 8B) . The levels of CTGF protein (36 kDa), however, were not altered within the selected perfusion period (Fig. 8) . 

Facility Studies

During the first 8 hours of the adaptation period, facility increased in both groups of eyes, before it slowly declined and finally reached a steady state level at t = −15 hours (Fig. 9A) . This corresponds to a phenomenon called “washout,” which was previously described for other, nonhuman perfusion systems. ²¹ Erickson-Lamy et al. ²¹ explained this as a result of ECM loss due to perfusion. The average facility during a time period of 10 hours before medium exchange (t = −10 hours to t = 0 hours) was 0.445 ± 0.007 (SD) μL/min per mm Hg in eyes of the TGF-β₂-perfusion group and 0.447 ± 0.011 μL/min per mm Hg in the control group, representing a pressure of 10.12 ± 0.16 mm Hg in the TGF-β₂-perfusion group and 10.07 ± 0.25 mm Hg in the control group. Within the control group, the facility declined slightly to 0.430 ± 0.005 μL/min per mm Hg within 60 hours after medium exchange (an average 10-hour time period from t = 50 hours to t = 60 hours), corresponding to a perfusion pressure of 10.46 ± 0.13 mm Hg (Fig. 9C) . In contrast, in TGF-β₂-perfused eyes facility decreased down to 0.268 ± 0.002 μL/min per mm Hg during the same period, representing a 37.7% reduction of facility compared with the control group (P < 0.001). Expressed as normalized data (Fig. 9B) , the facility decrease due to TGF-β₂ was 33.5% (P < 0.001). The resultant perfusion pressure was 16.78 ± 0.11 mm Hg in that group (Fig. 9C) . 

Discussion

The porcine perfusion model is frequently used to analyze the effects of putative factors involved in glaucoma pathogenesis by examining facility reduction and morphologic changes. Because porcine material can be obtained in almost unlimited amounts, biochemical and molecular analyses of samples in adequate statistical extents are possible. However, there are marked differences in the overall anatomy and morphology between porcine and primate eyes. ^{19 20} How these differences might affect the transferability of porcine data to the primate eye has not yet been sufficiently tested. The purpose of the present study therefore was to investigate the ultrastructural differences in the juxtacanalicular region in the porcine and primates TM and to test how the preparation technique and the perfusion itself affects the integrity of the outflow pathways. 

Our new ultrastructural investigations of the porcine cribriform region revealed a similar situation compared with primate TM. Studies of serial ultrathin sections through different planes of the primate subendothelial region have shown that the elastic network is connected to the endothelial cells of SC by fine fibrillar material splitting from the elastic sheaths (connecting fibers). ^{22 23} Recently, we found that the elastic cribriform network is also connected to the cribriform cells. ²⁴ Because the tendons of the CM insert into the elastic cribriform net, it was assumed that the cribriform cells are exposed to a mechanical tension, resulting in the expression of the stress protein αB-crystallin. ^{25 26}  

Obvious morphologic differences between porcine and primate TM were restricted to the inner TM and confirmed the descriptions in the literature. With respect to the juxtacanalicular and subendothelial region, we found that the porcine TM also contains an elastic network with a structure similar to that in the primate TM. The elastic fibers are surrounded by sheaths, which are connected to the cribriform TM cells and the endothelium of the aqueous plexus—the porcine analogue of SC. Through the connection to the SP and the CM, the elastic network and its connections to the cribriform cells and vessel endothelium are expanded, thereby presumably preventing the collapse of the loops. ²⁰ The fact that endothelial BM is lacking at places associated with optically empty spaces might indicate areas of fluid flow or fluid pathways. These areas are presumably smaller and filled with ground substance in situ, but that material is supposedly lost during embedding procedures. ^{6 27} In primate eyes, the use of other microscopic techniques resulted in a substantial size reduction of these optically empty spaces, ²⁸ confirming that assumption. These optically empty pathways in the porcine eyes ended at endothelial cells which formed giant vacuoles, further contributing to the hypothesis of their being fluid pathways. On the ultrastructural level, our data show that porcine cribriform cells have large amounts of cell organelles known to be involved in protein synthesis and secretion, like the rER. It is tempting to speculate that these cribriform cells produce and secrete ECM components or enzymes regulating its formation and degradation and thereby play a role in the regulation of outflow resistance. 

Morphologic analysis of perfused porcine eyes showed that changes were more or less completely restricted to the inner TM as a consequence of the preparation. With our modified preparation method, in which we left part of the CM in place, the cribriform region remained expanded between the SP and cornea and vessel loops and fluid pathways remained open. In contrast, when too much of the CM was excised, either the juxtacanalicular TM was collapsed, due to missing support by the CM, or most of the TM cells were washed out. In most of these preparations, physiological measurements resulted in ambiguous pressure curves and made subsequent biochemical experiments impossible. In contrast, when the entire CB was left in place, the TM cells were overloaded with pigment granules and a profound loss of cells was detectable. Perfusion curves then showed increased volatility similar to that found in perfused human eyes. ²⁹ From that we conclude, that although our modified preparation method has a good success rate, matching of the morphology with physiological or biochemical data is essential to obtain reliable results. 

Despite all the morphologic similarities of unperfused and perfused anterior eye segments, we found one significant difference—namely, the staining pattern of αB-crystallin. In the perfused eyes, this stress protein was expressed throughout almost the entire TM, whereas the unperfused control showed very little staining. As already mentioned, in the primate TM the cribriform cells express more αB-crystallin than the inner TM cells, presumably due to a more pronounced mechanical stress as the ciliary muscle tendons directly or indirectly insert through the cribriform elastic network into the BM of the cells. ^{25 26} In porcine eyes the elastic network of the TM is also connected to the TM cells, but there are no tendons, as are found in the primate eye, and the CM is less developed. ²⁰ Therefore, the mechanical tension on the cells may be less intense. In the perfused eyes, however, the support of the TM by the CM is changed as a consequence of the preparation. This could cause a higher mechanical stress, affecting the entire TM. A second possibility that could cause stress on the cells is the composition of the medium used for perfusion studies. Fautsch et al. ³⁰ have found that anterior segment perfusion with collected porcine aqueous humor has a different outcome than does perfusion with synthetic medium (e.g., myocilin formation is different). This observation indicates that there are unknown factors in the AH that are lacking in the perfusion medium, which could lead to the observed stress response. 

Although this is a deficiency in the use of medium, the defined composition of DMEM allows the controlled and thorough investigation of the effects of factors supplementing the medium. We studied the effects of TGF-β₂, because extensive data concerning this cytokine in the human perfusion model and cell cultures are available. Porcine anterior segments reacted with a decrease in outflow facility and an activation of TM cells, as human anterior segments did in the analogous experiments. ¹³ Our biochemical investigations revealed that porcine TM cells in situ respond to TGF-β₂ with increased expression of fibronectin and αB-crystallin, as do cultured human TM cells. ^{11 26} This finding also confirms the impression we had regarding immunohistochemical staining, where staining intensities of both proteins seemed augmented. However, morphologically, we could not demonstrate increased amounts of fibrillar material in the subendothelial region of the outflow vessels. We assume that this may be due to the short perfusion time and the mentioned problem of loss of extracellular material during fixation and embedding. However, we found significant increases in cell organelles (i.e., rER and Golgi apparatus), indicating a TGF-β₂ mediated activation of protein synthesis by the TM cells adjacent to the fluid pathways. Western blot analysis confirmed that expression of factors like PAI-1 and TSP-1 was indeed increased. The moderate inductions we measured are presumably also a consequence of the short perfusion time. 

Our data indicate that TGF-β₂ in the porcine TM activates protein synthesis in common, as shown by increased amounts of cell organelles, and specifically expression of FN and PAI-1, confirming the human data. ^{11 12} Thereby ECM formation may be promoted and degradation simultaneously inhibited, presumably resulting in the observed increase in outflow resistance. 

Taken together, the results of our study indicate that the porcine perfusion model represents an adequate substitute for the human perfusion system, with respect to morphologic and biochemical conditions. By the unlimited availability of material it allows extended physiological, morphologic, and molecular investigations of the effects of putative glaucoma-promoting factors on the TM, thereby providing new insights into the pathogenesis of this disease. 

²  Contributed equally to the work and therefore should be regarded as equivalent first authors.

 Supported by Grant SFB539 of the Deutsche Forschungsgemeinschaft.

 Submitted for publication October 26, 2005; revised December 22, 2005; accepted March 8, 2006.

 Disclosure: B. Bachmann, None; M. Birke, None; D. Kook, None; M. Eichhorn, None; E. Lütjen-Drecoll, None

 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.

 

Preparation of a porcine anterior eye segment for perfusion experiments (A–C) and excision of TM specimens (C, D). (A) Anterior chamber after removal of vitreous, choroid, and retina. Arrow: the incision position and cutting direction. (B) After incision, iris (I) and CB are lifted up and gently pulled outward with forceps (white arrows) while the disconnection from the subjacent TM ( ) is performed with fine scissors (black arrow). The remnants of the longitudinal portion of the CM and SP thereby remain in place. (C) After removal of I and CB, the TM ( ) is demarcated with a scalpel, posterior (p, dotted line) and adjacent to the pigmented SP and remnants of the anterior CM, and anterior (a, dotted line) at a distance of ∼2 mm toward the cornea (C). (D) The TM is excised with fine forceps as a complete band between the anterior (a) and posterior (p) border. L, lens; S, sclera; C, cornea.

Table 1.

 

Antibodies Used in the Study

Antibody	Abbreviation	Application	Dilution	Supplier
Cy3-conjugated mouse monoclonal anti-smooth muscle α-actin	mc m-α-α-smA-Cy3	IHC	1:250	Sigma
Rabbit polyclonal anti-αB-crystallin	pc r-α-αB-crystallin	IHC/WB	1:500/1:1,000	Stressgen
Rabbit polyclonal anti-fibronectin	pc r-α-FN	IHC/WB	1:200/1:750	Dako
Cy3-conjugated goat anti-rabbit	g-α-r-Cy3	IHC	1:2,000	Dianova
Alexa488-conjugated goat anti-rabbit	g-α-r-Alexa488	IHC	1:2,000	MoBiTec
Alkaline phosphatase–conjugated anti-digoxigenin	α-DIG-AP	NB	1:10,000	Roche
Mouse monoclonal anti-smA	mc m-α-smA	WB	1:500	Dako
Goat polyclonal anti-CTGF	pc g-α-CTGF	WB	1:500	Santa Cruz Biotech
Rabbit polyclonal anti-PAI-1	pc r-α-PAI-1	WB	1:250	Santa Cruz Biotech
Mouse monoclonal anti-TSP-1	mc m-α-TSP-1	WB	1:250	Biocarta
Alkaline phosphatase-conjugated donkey anti-goat	d-α-g-AP	WB	1:3,000	Santa Cruz Biotech
Alkaline phosphatase–conjugated donkey anti-rabbit	d-α-r-AP	WB	1:10,000	Santa Cruz Biotech
Alkaline phosphatase–conjugated rabbit anti-mouse	r-α-m-AP	WB	1:5,000	Dako

 IHC, immunohistochemistry; WB, Western blot; NB, Northern blot. Locations of suppliers: Sigma-Aldrich, Seetze, Germany; StressGen, Victoria, British Columbia, Canada; Dako, Glostrup, Denmark; Dianova, Hamburg, Germany; MoBiTec, Göttingen, Germany; Roche, Basel, Switzerland; Santa Cruz Biotechnology, Santa Cruz, CA, USA; Biocarta, Hamburg, Germany.

Figure 2.

 

Semithin sagittal sections through the temporal (A, C) and nasal (B, D) quadrants of the porcine chamber region before (A, B) and after 72 hours of medium perfusion (C, D). (A, B) In the temporal quadrant the scleral sulcus is shallower in the inward–outward direction than in the nasal quadrant, whereas its anterior–posterior diameter is longer temporally than nasally. In both regions, the outer TM is expanded between the SP and cornea (C), and loops of the aqueous vessels (VL) extend into the outer meshwork. In the temporal quadrant, part of the inner TM continues into the connective tissue of the CB, whereas nasally nearly the entire posterior meshwork inserts in the thick SP. (C, D) The outer meshwork is still expanded between SP and cornea (C), open vessel loops (VL) still extend into the outer TM. The trabecular cells are still present. The entire meshwork appears slightly collapsed. DM, Descemet’s membrane; I, iris.

Figure 3.

 

Electron micrographs of tangential sections through the subendothelial cribriform region of the vessel loops in the outermost TM (A) and the more lamellated region (B). (A) The cribriform elastic network (EL) shows the same morphology as its equivalent in the primate TM. The elastic fibers have an electron-dense central core intermingled with only a little elastin and are surrounded by a sheath of fine fibrils. In places (arrowheads) these fine fibrils of the elastic fiber sheath are connected to the cell membrane of the cribriform cells. The cribriform cells form elongated cell processes that are connected to each other or to the cell bodies of adjacent cells ( ). Within their cytoplasm, the cells contain numerous ribosomes (R) and membranes of rER. (B) The TM cells incompletely cover a connective tissue lamella, contain numerous profiles of rER and sections through cilia. BM, basement membrane; M, mitochondria. Magnification: (A) ×2273; (B) ×2237.

Figure 4.

 

Electron micrograph of an oblique section through the inner wall of a vessel loop extending into the outer TM. The cribriform elastic network (EL) located underneath the inner wall of the loops is connected to the inner wall endothelium (E) by fine fibrils splitting from the fiber sheath (arrowheads). The elastic fibers are also connected to the cribriform cells (CR; arrows). ( )An optically empty area underneath the inner wall endothelium next to basement membrane (BM)–free endothelial cells. M, mitochondria. Magnification, ×5449.

Figure 5.

 

Immunohistochemical staining for αB-crystallin (A–C) and FN (D–F) in sections of nonperfused (A, D), medium-perfused (B, E), and TGF-β₂-perfused porcine anterior eye segments (C, F). Frozen 10-μm-thick sagittal sections through the TM of the nasal region were stained for αB-crystallin (A–C) and FN (D–F). Staining for αB-crystallin was negative in the unperfused anterior segments (A), whereas both perfused eyes, showed staining of TM cells (B, C). Staining for FN was detectable throughout the entire TM (D–F). There were significant increased FN staining patterns after perfusion for 72 hours with 10 ng/mL TGF-β₂ (E, F).

Figure 6.

 

Electron micrographs of sagittal sections through (A) the inner wall endothelium (E) of a vessel loop and an adjacent cribriform (CR) cell and a TM cell of the lamellated region after 72 hour of perfusion with 10 ng/mL TGF-β₂. There were large cisterns of rER in the cribriform cells filled with an electron lucent material (A), as well as in the TM cells of the outer lamellated TM (B). Magnification: (A) ×2465; (B) ×2324.

Figure 7.

 

Northern blot analysis of αB-crystallin and FN mRNA expression (A) and quantification of data (B). Seventy-two hours of perfusion with 10 ng/mL TGF-β₂ induced expression of αB-crystallin and FN mRNA in porcine TM cells compared with the medium-perfused control (A, co). 28S and 18S rRNA were stained with methylene blue and served as the loading control. For quantification (B), signal intensities of the medium-perfused control were set to 1, and expression levels of TGF-β₂-perfused TM probes are given as x-fold of the control (mean ± SD, n = 4). Differences in 28S and 18S band intensities were considered for quantification.

Figure 8.

 

Western blot analysis of αB-crystallin, FN, PAI-1, TSP-1, and CTGF protein expression (A) and quantification of data (B). Seventy-two hours of perfusion with 10 ng/mL TGF-β₂ induced expression of αB-crystallin, FN, PAI-1, and TSP-1 in porcine TM cells compared with the medium-perfused control (A, co). Equal loading was assured by determination of protein contents before loading. For quantification (B), signal intensities of the medium-perfused control were set to 1, and expression levels of TGF-β₂-perfused TM probes are given as x-fold of the control (mean ± SD, n = 3). Protein band sizes of a molecular weight standard run concurrently are indicated.

Figure 9.

 

Plot of normalized outflow facility versus time (A), percentage facility (B), and resultant perfusion pressure (C). (A) Facility decreased within 24 hours after initiation of perfusion with 10 ng/mL TGF-β₂ (t = 50 hours). Facility gradually decreased for approximately 60 hours until reaching a steady state value (t = 110 hours) approximately 33% lower than that in the contralateral control eyes. These curves are an average of data from 30 pairs of porcine eyes, scaled so that the time of medium exchange with TGF-β₂ is t = 50 hours in all eyes. Normalized facility is instantaneous facility divided by the facility in the 24-hour period before exchange. (B) Mean average percentage facility (± SD; n = 30) in eyes just before anterior chamber exchange (baseline) and 72 hours later. In TGF-β₂-perfused eyes, facility decreased to 66.5% after 72 hours. The facility of medium-perfused control eyes before exchange was set to 100%. *Statistical significance (P < 0.001). (C) Resultant mean perfusion pressure (±SD; n = 30) calculated from the perfusion rate. Perfusion with TGF-β₂ increased pressure to ∼30 mm Hg. Pressure before medium exchange was 18 mm Hg in eyes designated for TGF-β₂ treatment and control eyes. Pressure in control eyes stayed constant at 18 mm Hg. *Statistical significance (P < 0.001).

The authors thank Ross Ethier for assistance with construction of the perfusion system and Anke Fischer, Katja Gedova, Marco Gösswein, Gerti Link, Julia Müller-Mausolf, and Heide Wiederschein for expert technical assistance.