Viscocanalostomy in Rhesus Monkeys FREE

Anterior chamber drainage surgery is an essential tool in glaucoma management.Nonpenetrating techniques are being refined to reduce the postoperative complicationsassociated with more traditional drainage procedures. Viscocanalostomy anda related procedure called deep sclerectomy havegarnered much attention recently. Although more difficult and perhaps slightlyless effective at lowering intraocular pressure (IOP) than standard trabeculectomy,¹ they lessen the risk of some of the postoperativeproblems of trabeculectomy and thus seem attractive.

Stegmann et al² recognized the need fora technique that would be successful in areas with a high risk of postoperativeinfection. They proposed the viscocanalostomy procedure, based on Krasnov’s³ and Zimmerman and coworkers’⁴ workon nonpenetrating “trabeculectomy.” Ideally, there is no obviousor intentional invasion of the anterior chamber and no iridectomy. These patientstypically have quiet eyes postoperatively, do not become hypotonous, and,in 1- to 5-year follow-up, have good IOP control, albeit not as low as withtrabeculectomy.^{2 ,5 - 7} Theyoften have no clinically visible filtering bleb, or they may have a small,relatively flat bleb, quite different from results with trabeculectomy. Antimetaboliteis not used.

Critical components of the surgical procedure are creation of superficialand deep scleral flaps, unroofing of Schlemm canal and Descemet membrane followedby cannulation of the cut ends of Schlemm canal and intracanalicular injectionof high-molecular-weight 1.4% sodium hyaluronate (Healon GV; Pfizer Inc, NewYork, NY), and excision of the deep scleral flap. The hypothesis was thataqueous humor leaves the anterior chamber by percolating through Schlemm canalendothelium and Descemet membrane into the hollowed-out “lake,”and then enters the widened cut ends of Schlemm canal.

As pointed out by Johnson and Johnson,^{8 - 9} itis unclear how this mechanism would be effective in lowering outflow resistance,as the current concept of aqueous outflow dynamics implies that most of theresistance to outflow of aqueous humor is internal to Schlemm canal endothelium.¹⁰ Indeed, more recent histopathological studies onhuman and monkey cadaver eyes showed that the dilation of Schlemm canal withviscoelastic agents (1.4% sodium hyaluronate or 2.3% sodium hyaluronate [Healon5; Pfizer Inc]) caused disruptions in the endothelial lining of Schlemm canal,^{11 - 12} an effect that could significantlyreduce outflow resistance in the trabecular meshwork (TM), but would implythat viscocanalostomy is a penetrating rather than a “nonpenetrating”surgical technique. To obtain more data on the functional and structural effectsof viscocanalostomy in the living eye, we performed viscocanalostomy in theeyes of rhesus monkeys and studied structure and function of the outflow pathways1 to 2 months after surgery.

Four normal adult male and female rhesus monkeys (Macaca mulatta), aged 19 to 20 years, were studied. Baseline slitlampbiomicroscopy and IOP measurements using a minified Goldmann applanation tonometer¹³ were performed. Baseline perfusion outflow facilityand tonographic outflow facility were determined on 2 animals (monkeys AI09and AI10). Perfusion outflow facility was determined by 2-level constant-pressureperfusion of the anterior chamber with the use of Bárány solution,¹⁴ correcting for internal apparatus resistance.¹⁵ The eyes were cannulated with a branched needle,with one branch connected to a pressure transducer and the other branch toa reservoir inflow line. Tonographic outflow facility was determined withan electronic Schiøtz tonographer (model 720 D; Berkeley Bioengineering,San Leandro, Calif). Applanation tonometry was done before tonography. IfIOP readings were less than 20 mm Hg, the 5.5-g weight was used; between 20and 29 mm Hg, the 7.5-g weight was used. Tonography was done for 4 minuteson each eye. Facility was calculated by means of tables and the Friedenwaldnomogram, according to methods described in Becker-Shaffer’sDiagnosis and Therapy of the Glaucomas.¹⁶ Allanimals had IOP measured weekly with the Goldmann tonometer. Viscocanalostomysurgery (see description in the “Surgical Technique” section)was performed after perfusion-induced inflammation had subsided (28 days).Two monkeys (81051 and AI45) were used as practice surgery animals. Theseanimals were not originally intended to be part of the study and so did nothave baseline facility measurements performed. They were to be used to workout any technique differences or problems we might encounter in adapting ahuman procedure to monkeys. The surgery went well, and these animals wereincluded in subsequent testing. In all, surgery was done in 1 eye of eachof 4 animals. None of the animals received a sham procedure. One practiceanimal (AI45) had a structurally abnormal control eye unsuitable for physiologicaltesting. The monkeys were divided into 2 groups, each with 1 normal and 1practice animal. Group 1 had perfusion and tonographic outflow facility determinationsand group 2 had tonographic outflow facility determination 33 to 42 days aftersurgery. The animals were divided into 2 groups to permit different structuralapproaches and physiology protocols for maximum information.

After perfusion inflammation had subsided (21 to 28 days after perfusion,63 days after surgery), both anterior chambers of each monkey were exchangedwith cationic 5-nm and noncationized 10-nm gold solution as tracers to delineateflow pathways and to label extracellular matrix, at an IOP of approximately15 mm Hg, and then perfused at 25 mm Hg with Ito solution from an elevatedreservoir. Under deep general anesthesia with intravenous pentobarbital sodium,15 mg/kg, these animals were then perfused through the heart with phosphate-bufferedsaline, 0.1 mol/L (pH 7.4), followed by Ito solution. The eyes were enucleated,windows were cut in the cornea and sclera, and the eyes were placed in thesame fixative and sent to Germany for electron microscopy.

On arrival, the eyes were placed in cacodylate buffer (pH 7.4) for 24hours to wash out fixative. Each eye was bisected and the anterior halveswere cut into quadrants by meridional sectioning. Each quadrant was furtherdissected into wedge-shaped specimens 1.0 to 1.2 mm wide that contained TM,ciliary muscle, iris, and adjacent cornea and sclera. In the superior quadrant,the distance of each specimen to the 12-o’clock limbus, which had beenmarked by a suture during enucleation of the eye, was identified and documented.All wedges were dehydrated in ascending concentrations of alcohol and embeddedin epoxy resin according to standard protocols. One-micrometer semithin sectionswere cut from each specimen of the superior quadrant (6-8 per monkey) andfrom at least 3 specimens of the other 3 quadrants. All semithin sectionswere stained with toluidine blue O and examined by light microscopy. Subsequently,ultrathin sections were cut from each specimen that had been investigatedby light microscopy and stained with lead citrate and uranyl acetate. Bothsemithin and ultrathin sections were assigned a unique identifying numberand were examined by a masked observer (E.R.T.).

Tonographic outflow facility determination at day 35 after viscocanalostomywas followed by perfusion through the heart with phosphate-buffered neutralformalin on day 36. The eyes were enucleated and placed in 10% formalin forprocessing and analysis by one of us (D.M.A.). The eyes were histologicallyprocessed, embedded in paraffin, and serially sectioned at 5 μm. Every10th section was stained with hematoxylin-eosin and coverslipped. The eyesections were assigned a unique identifying number and a masked observer examinedthe hematoxylineosin–stained sections under a microscope. The eyes wereexamined and several histologic features were recorded: (1) the patency andintegrity of Schlemm canal, (2) evidence of breaks in the TM, (3) the presenceof scleral lakes, (4) the presence and degree of inflammation, (5) the presenceand degree of fibrosis, and (6) any other unusual histologic ocular features.Emphasis was placed on the area of, and adjacent to, the surgical site, aswell as the area 180° from the surgical site. The surgically treated eyesfrom group 2 monkeys were compared with the contralateral control eye fromone of the animals.

Intramuscular ketamine hydrochloride anesthesia (10 mg/kg, supplementedevery 20-30 minutes as needed with 5 mg/kg) was used for all procedures. Inaddition, animals received intravenous pentobarbital sodium (10-15 mg/kg)for perfusion outflow facility, or acepromazine maleate (0.2-0.5 mg/kg intramuscularly)for tonographic outflow facility. For perfusion, animals were maintained underketamine anesthesia until slitlamp biomicroscopy and IOP measurement wereperformed, intravenous lines were in place, and the perfusion apparatus wascalibrated. At that point they were given pentobarbital and no further ketamine.For the viscocanalostomy surgery, animals were intubated and maintained underisoflurane inhalation anesthesia. All were maintained with intravenous fluids(5% dextrose–lactated Ringer solution) at a rate of 10 mL/kg per hour.Animals were given subtenon gentamicin sulfate and methylprednisolone acetate(20 mg) postoperatively. Penicillin G procaine and penicillin G benzathine,50 000 U/kg, were given intramuscularly for 5 days and methylprednisoloneacetate, 1 mg/kg, was given intramuscularly for 4 weeks, with the dose tapereddown during the last week.

All experiments were performed in accordance with the Association forResearch in Vision and Ophthalmology Statement on the Use of Animals in Ophthalmicand Vision Research.

After a bridle suture was passed under the superior rectus muscle, afornix-based conjunctival flap was prepared. To avoid damage to Schlemm canal,collector channels, and the sclera itself, hemostasis was maintained by repeatedirrigation with ornipressin solution, 5 IU/mL (Por 8; Sandoz, Basel, Switzerland),so as to use as little thermal coagulation as possible.

With a 20-gauge diamond knife, a 5×5-mm limbus-based parabolicincision was made in the sclera, and a 200-μm superficial flap was dissectedwith a single-use bevel-up spatula (both from Grieshaber, Schaffhausen, Switzerland).By the same technique, an inner concentric 4×4-mm limbus-based scleralflap was sculpted beneath the previous one, keeping the surface of the cutso close to the choroid as to have a dark reflex. When the cut was advancedlimbally, Schlemm canal was deroofed and the 2 openings of the canal remainedpatent at the edges of the cut.

The ostia of Schlemm canal were then cannulated with a 190-μm-diameterblunt cannula (Grieshaber) and filled with high-molecular-weight 1.4% sodiumhyaluronate. To limit damage to the Schlemm canal walls, the injection ofsodium hyaluronate was started while the ostia were approached, the cannulainsertion did not exceed 1 mm, and a small amount of sodium hyaluronate wasslowly injected on each side 3 to 4 times. Approximately 275 μL of sodiumhyaluronate was injected. Most of it did not go into Schlemm canal becauseof high reflux.

By gently pulling the inner scleral flap upward and delicately depressingthe floor of the canal and Descemet membrane with the tip of a cotton swab,the membrane itself was cleaved anteriorly from the cornea for approximately1 mm (Figure 1), and aqueous was seento percolate through this window and to enter the lake. As soon as the windowwas completed, the inner scleral flap was excised by means of Vannas scissors.

The ostia of Schlemm canal were cannulated again and sodium hyaluronatewas gently injected 2 to 3 times in each opening. Finally, the outer flapwas tightly sutured with seven 10-0 nylon stitches, and sodium hyaluronatewas injected underneath the flap to temporarily fill the intrascleral lakeand prevent it from collapsing and scarring in the early postoperative period.Finally, 2 wing 8-0 silk sutures were passed to hold the conjunctiva in place.

In a separate experiment, approximately 25 to 30 μL of sodium hyaluronatewas injected into the anterior chamber of 1 eye of an adult female cynomolgusmonkey (Macaca fascicularis). Approximately 25 to30 μL of aqueous humor was then removed and the IOP was checked to ensurethat it was not elevated. Slitlamp examination was performed 24 hours laterto assess the amount of inflammation, which was minimal. Fixation for thisanimal was similar to that of the animals in group 1. Under deep general anesthesiawith intravenous pentobarbital sodium, 15 mg/kg, the animal was perfused throughthe heart with phosphate-buffered saline, 0.1 mol/L (pH 7.4), followed byIto solution. The eyes were enucleated, windows were cut in the cornea andsclera, and the eyes were placed in the same fixative and sent to Germanyfor electron microscopy. Wedge-shaped specimens from each quadrant of theanterior eye were processed for light and electron microscopy as describedin “Tissue Fixation, Group 1.”

One animal had preoperative and postoperative perfusion outflow facilitydeterminations (monkey AI10). Postoperative facility (35 days after surgery)increased by 50% in the treated eye and 22% in the control eye of this animalcompared with baseline. The ratio of facility in the viscocanalostomy-treatedand control eyes (V/C) changed from 0.78 before to 1.22 after surgery. Tonographicoutflow facility was determined before and 34 days after surgery on 2 animals(monkeys AI09 and AI10). Facility increased by 68.5% ± 0.5%in the viscocanalostomy-treated eyes and 42.5% ± 0.5% inthe control eyes. (Results are given as mean ± SEM unlessotherwise indicated.) The V/C facility ratio changed from 0.66 ± 0.10before to 1.21 ± 0.16 after surgery. Tonographic outflowfacility data were collected 1 day before perfusion outflow facility on 3occasions and 9 days before in 1 other. One animal had both preoperative andpostoperative testing done (AI10) and so accounted for 2 of the 4 test points.The ratios of V/C facility were calculated for both types of facility measurement.The ratio of V/C facility for tonographic outflow was 0.86 ± 0.18.The ratio of V/C facility for perfusion outflow was 0.84 ± 0.12.The ratio of tonographic facility to perfusion facility was 1.00 ± 0.07.

The presurgery IOP for the 4 eyes that underwent the viscocanalostomyprocedure was 17.3 ± 1.6 mm Hg. The IOP at time of deathfor these eyes was 16.3 ± 0.6 mm Hg, a nonsignificant decreaseof 6%. The presurgery IOP for the 3 fellow eyes was 17.3 ± 1.2mm Hg, while the IOP at time of death was 16.7 ± 0.3 mm Hg,a nonsignificant decrease of 4%.

The eyes of 2 monkeys that underwent viscocanalostomy in 1 eye wereexamined at 63 days after surgery. The entire circumference of operated-onand control eyes was investigated. In the operated-on eyes of both animals,Schlemm canal and juxtacanalicular TM had disappeared in the center of thesuperior quadrant, where the operation had been performed (Figure 2). In an area of about 2 mm in circumferential width, irregularwhorls of elastic and collagenous fibers occupied the position where Schlemmcanal originally had been located (Figure 2).At the inner aspect of the TM, uveal trabecular lamellae covered by TM cellsremained intact and the chamber angle was open.

Adjacent to the operated-on area, Schlemm canal was present (Figure 3A) in both operated-on eyes. In noneof the eyes was Schlemm canal abnormally dilated. There were, however, markedstructural changes in the outer parts of the TM along a distance of about2 to 6 mm next to the operated-on site. In this area, trabecular lamellaehad become replaced by a network of fibroblastlike cells that were embeddedin a loose collagenous matrix (Figure 3A).This tissue appeared not to impede flow of aqueous humor, as giant vacuoleswere present along the inner wall of Schlemm canal (Figure 3A). In places, the juxtacanalicular zone was filled withhomogeneous material that was not seen in control eyes. By electron microscopy,this material was electron dense and had a fine granular structure (Figure 3B). The inner wall endothelial liningof Schlemm canal frequently showed intercellular openings approximately 100nm to 2 μm in diameter (Figure 3Band C). Through these openings, extensions of the fine granular material protrudedinto the lumen of Schlemm canal. Both 5- and 10-nm gold particles that wereembedded in the fine granular material were seen at the outer side of theinner wall next to the openings, as well as in the part of the material thatextended through the openings (Figure 3C).We hypothesized that the fine granular material was sodium hyaluronate thathad been injected into Schlemm canal during surgery and was pressed throughthe openings in Schlemm canal endothelium into the juxtacanalicular regionof the TM. To support this hypothesis, we performed an additional experimentand injected sodium hyaluronate directly into the anterior chamber of theeye of an additional monkey. Twenty-four hours after injection, sodium hyaluronatewas observed by light microscopy in contact with the surfaces of ciliary processesand iris as homogeneous material that stained intensely with toluidine blueO (Figure 4A). By electron microscopy(Figure 4B), the same material was electrondense and finely granular, and showed essentially the same ultrastructuralcharacteristics as the material that was found in the juxtacanalicular TMof monkeys that underwent viscocanalostomy. Similar electron-dense and finelygranular material was observed after anterior chamber injection at the innerside of the peripheral cornea (not shown) and at the inner surface and theintertrabecular spaces of the uveal meshwork (Figure 4C and D), but not in the juxtacanalicular region.

In one of the viscocanalostomy-treated monkeys (AI10), at a distanceof about 2 mm from the 12-o’clock limbus, a large triangular defectwas present in the sclera adjacent to the anterior portion of the ciliarymuscle (Figure 5A). The defect had acircumferential extension of about 1 mm. The walls of the defect had a lengthof approximately 0.5 to 0.6 mm and were not covered by cells. We concludedthat this defect resulted from the operation and was a remaining part of theintrascleral reservoir, the so-called scleral lake that was created by removingthe inner layer of the sclera. Between the scleral lake and the TM, the scleraappeared to be markedly less dense than in other parts of the eye (Figure 5A). In the TM of this region, there werefewer trabecular lamellae, and large intertrabecular spaces were present (Figure 5B). The lumen of Schlemm canal formedprotrusions toward the TM (Figure 5Cand D). Similar protrusions were not observed in control eyes. On serial sections,it became evident that these protrusions communicated directly with the intertrabecularspaces and were not Sondermann canals, which are blind diverticula in theinner wall of Schlemm canal that are completely covered by endothelium. Sondermanncanals are extremely rare in monkeys (E.R.T., unpublished data, 2003), consistentwith the findings in control eyes. A similar large scleral defect was notpresent in the same area of the other operated-on eye. However, in some distinctareas, larger defects, approximately 1 to 5 μm in length, in the endothelialcovering of Schlemm canal were observed that were not associated with thefine granular material nor with thrombocytes (Figure 6). Such defects were seen along the inner wall and, morerarely, in the outer wall of Schlemm canal in this area (Figure 6A and B).By electron microscopy, fine fibrillar material and sheath-derived plaquematerial was seen to partly bridge over the defects (Figure 6C). Numerous 5- and 10-nm gold particles were attached tothis extracellular material (Figure 6D).In a region 2 to 4 mm nasal from the 12-o’clock limbus and in an areawhere defects in the outer wall were frequently observed, the extracellularmaterial between the outer wall of Schlemm canal and the outer side of thesclera and cornea was markedly less dense than in the control eye or in otherregions of this eye (Figure 7). Thus,at the limbus, a sharp boundary was formed at the anterior end of the TM thatseparated corneal stroma of normal extracellular matrix density localizedanteriorly to the TM from stroma of considerably less density that was localizedopposite to Schlemm canal and TM. By electron microscopy of this area of diminishedextracellular matrix density, large electron-empty spaces were seen betweencells that showed typical ultrastructural characteristics of fibroblasts.The amount of collagenous fibers appeared to be greatly reduced.

At the nasal, temporal, and inferior sides of both eyes, approximately90° to 180° from the center of the surgical site and beyond the extentof the Schlemm canal cannulation, obvious scarlike changes in the TM werenot observed. Still, the normal contour of Schlemm canal had changed and oftenshowed irregularly branched cul-de-sac–like protrusions that reachedtoward the inner parts of the TM (Figure 8).Such protrusions were not observed in control eyes. The inner wall in thisarea formed numerous giant vacuoles, and the juxtacanalicular area was againpartly filled with homogeneous material. Along the inner wall, aggregationsof thrombocytes were frequently observed that appeared to cover defects inthe endothelial lining of the canal. By electron microscopy, individual thrombocyteswere found at the outer side of the inner wall in close association with intercellularjunctions of the endothelium, some of which appeared open (ie, the cells wereseparated), others closed (ie, the cells were not separated). In addition,aggregates of thrombocytes were observed (Figure9). In these areas, cytoplasmic protrusions of individual thrombocyteswere seen that filled the gaps between adjacent inner wall cells. Gold particleshad become attached to extracellular fibers close to the openings, indicatingthat this region was used for flow of aqueous. In addition, gold particleswere seen in intracellular vesicles in neighboring Schlemm canal endothelialcells next to the openings. In contrast, no gold particles were observed inassociation with the extracellular matrix in immediately adjacent areas thatwere separated from the lumen of Schlemm canal by an intact endothelial layer.More rarely observed were contracted thrombocytes that closed openings ofthe inner wall, or larger thrombotic plaques that consisted of numerous aggregatedthrombocytes and fibrinlike material between them (Figure 9D and E).

Discontinuities in the periphery of Descemet membrane were observedat the base of the opercular area, an extension of Descemet covering the anteriorTM, which is not present in humans (Figure 10). Between the ends of the discontinuous Descemet membrane, cellsor smaller open spaces up to 5 μm in width were present. In other areas,Descemet membrane was interrupted by a small slit only. These discontinuitiesare normal for monkeys and were present in both control and experimental eyesand distributed equally around the circumference. No evidence of iatrogenicdisruptions of Descemet membrane was observed in any eye.

The viscocanalostomy-treated eyes of monkeys AI09 and AI45 were examinedat 36 days after surgery, along with the contralateral control eye of monkeyAI09. In all viscocanalostomy-treated eyes, the sclera of the surgical areawas mildly edematous and contained at least minimal fibrosis of the scleralflap interface. Schlemm canal was not dilated in any eyes. The lamella ofthe corneoscleral and uveal TM in viscocanalostomy-treated eyes showed breaksthat were not observed in control eyes. There were microscopic breaks in boththe inner and outer wall of Schlemm canal in both viscocanalostomy-treatedeyes. In these 2 eyes, there were also breaks in the inner and outer wallsof Schlemm canal 180° from the surgical site. A full-thickness discontinuityin Descemet membrane of the peripheral cornea was seen in 1 animal (AI09).

Baseline physiological testing showed small differences in outflow facilitybetween the eyes for each animal; the eye with lower facility was selectedfor viscocanalostomy surgery. As with clinical studies,^{1 - 2 ,6 - 7 ,17} therewas a marked decrease in IOP after surgery. Still, in our monkeys this effectwas transient and recovery to near baseline occurred within 30 days. Outflowfacility, determined after surgery when the eyes were quiet, increased toa greater extent, relative to baseline values, in the eyes undergoing viscocanalostomy.By this time IOP had recovered to baseline levels. We assume that larger decreasesin IOP as reported for human patients with glaucoma and larger increases infacility were not observed because preoperative IOP and baseline facilityin our monkeys were within the reference range. In addition, the higher-than-physiologicpressure gradient or flow rate during external perfusion or tonography mightcontribute to the increase in outflow facility. During perfusion from an openreservoir, IOP can be up to 12 mm Hg higher than spontaneous IOP. The resultanthigher pressure gradient and flow rate across the TM could further loosenalready weakened (by the injection and retention of viscoelastic material)cell adhesions to each other and to their extracellular matrix in the innerwall and juxtacanalicular zone, relaxing the meshwork so as to facilitateflow through it,^{18 - 19} as wellas driving fluid through the inner wall breaks in Schlemm canal. Under thisscenario, the surgical procedure might be more clearly effective functionallyin patients with glaucoma with elevated IOP, especially if they are not receivingsecretory suppressants.

Experimental glaucoma can be induced in monkeys by laser treatment ofthe chamber angle tissues.^{20 - 21} However,this procedure induces scarring of TM and Schlemm canal^{22 - 23} andwould preclude cannulation and injection of sodium hyaluronate in Schlemmcanal as required for effective viscocanalostomy.

In contrast to the modest changes in outflow facility, structural changesof the conventional outflow pathways that persisted for at least 1 to 2 monthspostoperatively were more pronounced. These changes included numerous breaksin the inner and outer walls of Schlemm canal endothelium that were observedover large parts of the circumference in all of the operated-on monkey eyes.Gold particles with a high affinity for extracellular matrix that were addedto the perfusion fluid were observed in areas of disruption where underlyingextracellular matrix bridged the defects or was partly protruding into thelumen of Schlemm canal, and were directly associated with this matrix. Becausethese particles must have been carried by aqueous flow, we assume that aqueoushumor passed through the breaks in the endothelial lining of Schlemm canal.Comparable breaks were reported in a recent study on human and monkey cadavereyes that had undergone viscocanalostomy.¹¹ Inthese acute experiments, the walls of Schlemm canal became disrupted aftermarked dilation of Schlemm canal due to cannulation and injection of sodiumhyaluronate. It seems reasonable to assume that a similar scenario is responsiblefor the structural defects in Schlemm canal endothelium in our material. Still,dilation of Schlemm canal after injection of sodium hyaluronate appears tobe an acute and transient phenomenon, since in our material 1 to 2 monthsafter surgery, Schlemm canal was of normal size or appeared narrowed, or wascompletely obliterated close to the site of surgery. In addition, we did notobserve signs of disrupted septae bridging the lumen of Schlemm canal thatwere seen in acute experiments on cadaver eyes.¹¹

It is well established that acute disruption of the endothelial liningof Schlemm canal in monkey eyes, eg, by treatment with disodium EDTA, ethyleneglycol-bis(2-aminoethylether)-N,N,N′,N′-tetraacetic acid (EGTA),^{24 - 25} or cytochalasin B,^{24 ,26} isassociated with a marked increase in outflow facility. It has been calculatedthat the inner wall endothelium of Schlemm canal accounts for only less than10% of total outflow resistance,²⁷ since itforms numerous micrometer-size pores that allow relatively free passage ofaqueous humor.¹⁰ Therefore, the effects ofinner wall disruption on outflow facility are thought to be due to a washoutof extracellular matrix components from the juxtacanalicular or cribriformregion of the TM into the lumen of Schlemm canal. Data from microcannulationexperiments indicate that most of the outflow resistance is located in thisregion, at a distance of approximately 7 to 14 μm from the inner wallendothelium.²⁸ Notwithstanding this evidence,there is still uncertainty as to the exact site and nature of the resistance,with most advocating the juxtacanalicular region but some advocating the innerwall endothelium.²⁹ It seems reasonable toassume that focal disruption of the inner wall endothelium and resulting washoutof extracellular material from the outer parts of the TM caused the effectson outflow facility in our monkeys. Our findings are informative regardlessof the resistance distribution between the sites in question; indeed, fromthe viscocanalostomy mechanistic standpoint, both the extracellular matrixwashout and inner wall endothelial defects are important.

It is surprising that these defects remained open for at least 1 to2 months after surgery. In the living eye, defects in Schlemm canal endotheliumthat are larger than the physiologic pore size of 0.25 μm to approximately2 μm are occluded by platelet aggregation, a process that occurs withinminutes.³⁰ This process appeared to be ineffective,or at least significantly delayed, in our material, as only some pores wereobserved that were completely or partially occluded by thrombocytes. Moreover,it is unclear whether the thrombocytes were effective in sealing the endothelialdefects, as numerous gold particles were often seen associated with extracellularmatrix components close to occluded pores, but not in immediately adjacentareas that were separated from the lumen of Schlemm canal by an intact endotheliallayer, indicating that passage of fluid through the pores did still occur.A likely explanation for the persistence of endothelial defects in Schlemmcanal could be a direct or indirect action of sodium hyaluronate on thrombocyteaggregation. Indeed, we observed in some of the defects homogeneous, granular,electron-dense material that was not observed in control eyes. Although wedo not have direct molecular proof, we assume that this material reflectsthe presence of remaining sodium hyaluronate, since sodium hyaluronate thatwas directly injected into the anterior chamber of a monkey in a parallelexperiment showed similar ultrastructural characteristics. After anteriorchamber injection, sodium hyaluronate was found in the posterior chamber,on the inner surface of the peripheral cornea, and on the inner uveal partsof the TM, but not in the juxtacanalicular region. This distribution indicatesthat in a structurally intact TM, sodium hyaluronate is too viscous to enterthe fluid pathways of the juxtacanalicular region, which are considerablysmaller than those of the corneoscleral and uveal TM.¹⁰ Hyaluronanhas been shown to inhibit platelet adhesion and aggregation,³¹ andsodium hyaluronate might have similar effects on the adhesion of plateletsto Schlemm canal endothelium. Identification of the molecular processes bywhich sodium hyaluronate might prevent healing of Schlemm canal defects couldprovide important information on how to increase the effectiveness of agentsthat may be used to disrupt the endothelial lining of Schlemm canal to therapeuticallydecrease outflow resistance.

Reduction of IOP in viscocanalostomy putatively requires aqueous humorto percolate into an intrascleral reservoir (the so-called scleral lake) thatis created by removing the inner layers of the sclera. From there it is thoughtto enter the widened cut ends of Schlemm canal and/or cut ends of collectorchannels. As a scleral lake was created during viscocanalostomy of our monkeys,such a fluid pathway may have existed for a time but was absent by the timeof our investigation 1 to 2 months after surgery. At the site of surgery,Schlemm canal was obliterated and open ends were not observed. In addition,a larger open intrascleral reservoir next to Descemet membrane or Schlemmcanal endothelium did not remain in any of the monkeys. There were, however,regions of hydrated sclera next to Schlemm canal and close to the site ofsurgery in some of the monkeys. These regions were filled with cells expressingthe typical structural characteristics of scleral fibroblasts and looselyarranged collagen fibers, and might represent healing stages of a former sclerallake, based on surgical anatomy and postoperative ultrasound biomicroscopystudies in humans. In one of the monkeys, there was an intrascleral open spaceat a site considerably distant from Schlemm canal, which was likely a remainingpart of the scleral lake created by removing scleral tissue. Thus, scleralspaces that were generated by removing the inner scleral layers were largelyoccluded 1 to 2 months after surgery. However, the hydrated sclera indicatesthat fluid was still moving through this region, which may constitute a low-resistancepathway for egress of aqueous humor from the anterior chamber, and could alsobe functionally equivalent to a scleral lake. Ultrasound biomicroscopy dataon human patients indicate that clear scleral lakes may remain open untilat least 7 to 9 months after surgery.³² Thismight indicate that scarring and/or endogenous removal of sodium hyaluronateoccurs at a much faster rate in monkeys than in humans. Reasons for this mightbe species-specific differences or the presence of more active tissue repairmechanisms in monkeys vs human patients with glaucoma. On the basis of ourdata, we cannot say whether collector channels were cut open during surgeryand whether they remained so despite the ongoing healing process. Theoretically,an opening of collector channels could have had effects on outflow facility,as about 25% of outflow resistance appears to be localized distal to Schlemmcanal, probably in the aqueous veins.³³ Wedid not see cyclodialysis in these animals despite step serial sectioningin 2 different laboratories, in all probability excluding this as a possiblefactor for changes in outflow resistance in our experiments.

We did see discontinuities in the periphery of Descemet membrane inthe opercular area (an extension of Descemet membrane covering the anteriorTM, not present in humans), but these are normal for monkeys,³⁴ werepresent in control eyes as well, and were distributed equally around the circumference.These are not related to the surgical procedure, and, in fact, we saw no evidenceof iatrogenic disruptions of Descemet membrane in any eye. Because in bothhumans and normal monkeys aqueous humor can be seen percolating through Descemetmembrane during viscocanalostomy surgery, this implies either that microperforationsoccur but were missed in our examination because of their infrequency, orthat Descemet membrane and the endothelium at the corneal periphery are infact leaky in the absence of covering stroma. The latter seems especiallylikely in the monkey, given the normal discontinuities in Descemet membrane.This could be present but more subtle in the human. Indeed, small fissuresthat are frequently associated with Hassall-Henle warts have been describedin Descemet membrane in human peripheral cornea. These fissures are predominantlylocalized on the endothelial side of Descemet membrane and may contain processesof endothelial cells or collagen fibrils.^{35 - 36} Someof these fissures have been observed to penetrate the entire thickness ofDescemet membrane to the corneal stroma.³⁶

In summary, the most likely explanation for the decrease in outflowresistance in monkeys after viscocanalostomy is the focal disruption of theinner wall endothelium and the opening of the juxtacanalicular or cribriformregion of the TM. We hypothesize that a similar effect is the major reasonfor the decrease in IOP in human patients after this type of glaucoma surgery.Opening of the outer TM had only modest effects on facility in normal monkeysbut might be very effective in human patients, where the cribriform regionis the tissue that is most affected by the pathologic changes that occur duringprimary open-angle glaucoma.³⁷

Correspondence: Paul L. Kaufman, MD, Departmentof Ophthalmology and Visual Sciences, University of Wisconsin, 600 HighlandAve, Madison, WI 53792-3220.

Submitted for Publication: November 5, 2003;final revision received March 18, 2004; accepted May 14, 2004.

Financial Disclosure: None.

Funding/Support: This study was supported bygrant EY02698 from the National Eye Institute, Bethesda, Md (Dr Kaufman);grant SFB 539 from the Deutsche Forschungsgemeinschaft, Bonn, Germany (DrTamm); Pharmacia Corp (acquired by Pfizer Inc, New York, NY); Research toPrevent Blindness Inc, New York (Drs Kaufman and Albert); and the Ocular PhysiologyResearch and Education Foundation, Madison, Wis (Dr Kaufman).

Acknowledgment: We thank Karin Göhler,MTLA, for excellent technical assistance and Marco Gößwein forhis excellent processing of the electron micrographs.